103071 5 January 2015 Working paper Firms and Value Chains in Southern Africa Jodie Keane The emergence of the mainstream literature on Global Value Chains (GVCs) is an acknowledgment of the increased importance of intra-firm trade in global trade, which is an unprecedented phenomenon. This trend is a result of the fragmentation of global production across countries, as they have become integrated within the global trading system through trade and finance linkages. This paper explores the policy implications of the new GVC literature for Southern Africa, including with regard to the expected, as well as actual, differences between intra- and extra-regional trade patterns amongst South African Customs Union (SACU) countries. The paper also explores firm-level performance to date, particularly with regard to the achievement (or not) of economic and social upgrading, or other upgrading as highlighted by the GVC literature, and relationship with end markets and buyers. Shaping policy for development odi.org Acknowledgements This working paper has benefitted from the guidance and oversight provided by Dr Chris Stevens, Senior Research Associate, regarding the approach to trade data analysis. Jane Kennan, Research Officer has provided all trade data for analysis. Thanks to Amir Ali Motahari for research assistance with analysis of the Global Enterprise Surveys and literature review. Useful comments received by Dr Mareike Meyn have been received and incorporated. We are very grateful to Takahiro Fukunishi for his insights and constructive comments on the quantitative methodology. Finally, it has been a pleasure to work with, and we are grateful for the constructive comments received by, Thomas Farole, World Bank (the funder) on the content and research approach. Any errors remain the authors. Table of contents Acknowledgements ii List of figures and tables iv Executive Summary vi The new wave of GVC literature 1 1.1 Why GVCs matter for regions, countries and firms 2 1.2 GVCs in SSA and SACU 4 1.3 Upgrading: Why it matters, and link to GVC Governance 5 1.4 The link between regional and global value chains 9 1.5 Concluding Remarks 11 Methodology: Firm-level evidence on value chain participation 13 2.1 Use of Enterprise Surveys 13 Extra and intra-regional trading patterns 23 3.1 Products Produced by Exporting Firms 23 3.2 Intra-Regional Products 24 3.3 Firm-Level Characteristics 27 3.3.1 SACU Exporters and Non-Exporters 27 3.3.2 Comparison between Extra- and Intra-Regional Exporters 29 Analysis of enterprise surveys 31 4.1 Methodology 31 4.2 Domestic and Export Orientated SACU Firms 33 4.3 Intra- and Extra-Regional SACU Exporters 35 4.4 SACU and Other African Exporters 37 4.5 Comparison of SACU with Global Enterprise Surveys 38 Conclusions 42 References 44 Appendix 1 47 Appendix 2 50 Appendix 3: Summary of Enterprise Survey Analysis 61 Appendix 4 Global Enterprise Survey Comparison 65 iii List of figures and tables Figure 1: Participation in global value chains ...................................................................................... 4 Figure A1: South Africa: top five imports (average 2011-13) ............................................................ 50 Figure A2: South Africa: top five exports (average 2011-13) ............................................................ 52 Figure A3: Botswana: top five imports (average 2010-12) ................................................................ 53 Figure A4: Botswana: top five exports (average 2010-12) ................................................................ 54 Figure A5: Namibia: top five imports (average 2011-13) ................................................................... 55 Figure A6: Namibia: top five exports (average 2011-13)................................................................... 56 Figure A7: Lesotho: top five imports (average 2008 –9) .................................................................... 57 Figure A8: Lesotho: top five exports (average 2008-9) ..................................................................... 58 Figure A9: Swaziland: top five imports (average 2005 –7) ................................................................. 59 Figure A10: Swaziland: top five exports (average 2005-7) ............................................................... 60 Table 1: Summary of Results: Odds Ratios ...................................................................................... viii Table 2: Share of Trade in Value Added by Region, 1995 and 2011 (percentage) ............................ 3 Table 3: Firms included in Enterprise Survey .................................................................................... 14 Table 4: Data Utilised and Availability ............................................................................................... 15 Table 5: Data Not Utilised .................................................................................................................. 16 Table 6: Availability of Indicators in the Global Enterprise Survey .................................................... 17 Table 7: First Product Produced by Firms that Export ...................................................................... 20 Table 8: Exporting Firms, and Major Products Traded on an Extra-Regional basis (50% threshold) ........................................................................................................................................... 23 Table 9: Exporting Firms, and Major Products Traded on an Extra-Regional basis (90% threshold) ........................................................................................................................................... 24 Table 10: Intra-Regional Exports (50% threshold) ............................................................................ 25 Table 11: Intra-Regional Exports (90% threshold) ............................................................................ 25 Table 12: Summary of enterprise data for Southern Africa ............................................................... 27 Table 13: Intra-regional/extra-regional exporters .............................................................................. 29 Table 14: Orientated to Domestic Market .......................................................................................... 34 Table 15: Export Orientated ............................................................................................................... 34 Table 16: Results of Logistic Regression for Intra-Regional Exporters ............................................. 35 Table 17: Results of Logistic Regression for Extra-Regional Exporters ........................................... 36 Table 18: SACU and African Exporters ............................................................................................. 37 Table 19: Results of Logistic Regression for Global Enterprise Surveys .......................................... 38 Table 20: Summary of Results: Odds Ratios .................................................................................... 39 Table 21: East Asia and Pacific ......................................................................................................... 40 Table 22: Eastern Europe Central Asia ............................................................................................. 40 Table 23: Latin America and Caribbean ............................................................................................ 41 Table 24: South Asia ......................................................................................................................... 41 Table A1: Review of Indicators of Interest ......................................................................................... 47 Table A2: South Africa: suppliers of top five imports (average 2011-13) .......................................... 51 Table A3: South Africa: markets for top five exports (average 2011-13) .......................................... 52 Table A4: Botswana: suppliers of top five imports (average 2010-12) .............................................. 53 Table A5: Botswana: markets for top five exports (average 2010-12) .............................................. 54 Table A6: Namibia: suppliers of top five imports (average 2011-13) ................................................ 55 Table A7: Namibia: markets for top five exports (average 2011-13) ................................................. 56 Table A8: Lesotho: suppliers of top five imports (average 2008 –9) .................................................. 57 Table A9: Lesotho: markets for top five exports (average 2008-9) ................................................... 58 Table A10: Swaziland: suppliers of top five imports (average 2005 –7) ............................................ 59 Table A11: Swaziland: markets for top five exports (average 2005 –7) ............................................. 60 iv Table A12: Ownership and Direct Exporting SACU Firms ................................................................ 61 Table A13: Firm Age SACU Sample ................................................................................................. 61 Table A14: Firm Size SACU Sample ................................................................................................. 61 Table A15: Licences and International Quality Certification .............................................................. 62 Table A16: Length of relationship with supplier, and sub-contracting ............................................... 62 Table A17: Manager’s experience in SACU firms ............................................................................. 62 Table A18: Managers’ education level in exporting firms .................................................................. 63 Table A19: Workforce skills and development .................................................................................. 63 Table A20: Workforce Education: An obstacle for SACU firms? ....................................................... 63 Table A21: Labour Market Regulations: An obstacle? ...................................................................... 63 Table A22: Customs and Trade Regulations: An Obstacle? ............................................................. 64 Table A23: Global Enterprise Survey Data ........................................................................................ 65 Table A24: Ownership and Exporting ................................................................................................ 65 Table A25: Firm Age: Global Enterprise Survey................................................................................ 66 Table A26: Firm Size Global Enterprise Survey Results ................................................................... 66 Table A27: Managers Experience ..................................................................................................... 66 Table A28: International Quality Certification Enterprise Survey ...................................................... 67 Table A29: Workforce Education: An obstacle? Regional Results................................................... 67 Table A30: Regional Results for Labour Market Regulations ........................................................... 67 Table A31: Regional Results for Customs and Trade Regulations ................................................... 68 v Executive Summary Despite lots of new descriptions of Global Value Chains (GVCs) and new databases, there remains rather more limited discussion as to what the implications of this increased phenomenon really entails. This is particularly so regarding discussions on the promotion of intra-regional trade. Some uncertainty stems from the clear tensions apparent within the new wave of GVC literature. On the one hand, more recent additions to the GVC case-study based literature suggest that intra-regional trade may be more amenable to some types of upgrading (such as moving from one function to another) compared to extra-regional markets (because marketing and retailing nodes are controlled by lead firms). On the other hand, given the degree of global fragmentation of production, it must be recognised that powerful new trade opportunities exist for countries that lack nascent industrial structures. This is because they can now specialise in a particular ‘task’ within a value chain , rather than a complete chain. As a result of engaging with lead firms, rapid economic and social upgrading at nodes of production may ensue, even if a change in the functional position of producers does not. The case-study based GVC literature emphasises how the relative merits of participating in one value chain compared to another essentially depend on how trade is managed and governed. Attention is drawn in this literature to the drivers of GVCs and the nature of relationships between firms that operate across borders; this is in addition to identification of the scope for governments to influence these structures so as to achieve upgrading. Although entering into GVCs may upgrade producers at a point in time and enable access to global markets, there is rather more limited evidence on how to sustain productivity growth overtime. Debates on the achievement of learning by exporting at the firm-level are still relevant. This discussion is amplified somewhat regarding discussion on the relative merits of the facilitation of intra- compared to extra-regional trade flows. In order to provide empirical evidence that either supports or rejects propositions regarding the relative merits of trading on an intra- or extra-regional basis in terms of firm-level upgrading trajectories, this paper develops a research methodology which integrates the World Bank Enterprise Surveys and trade data. Our research objective is to provide evidence on the relative merits of exporting on an intra- or extra-regional basis for firms in Southern Africa, and hence, to draw policy relevant implications. These include in relation to trade-related interventions, including those which may be operationalised at the intra-regional, rather than country level. We identify a number of indicators of interest within the World Bank Enterprise Survey which are relevant in terms of understanding how firms participate in GVCs. We move beyond the usual use of these surveys in terms of interpretation of doing business indicators across countries, and instead identify relevant indicators relating to entering and accessing GVCs. Our intention is to better integrate firm-level analysis into GVC analysis. Although input:output tables have provided new insights regarding country participation in GVCs, we have more limited information regarding the characteristics of firms involved. Overall, we present new evidence on firm-level heterogeneity across different types of value chains, including intra- and extra-regional value chains. This study is organised as follows. In Section 1 we summarise the new literature on GVCs and refer to evidence on the degree of participation in Southern Africa with vertically fragmented trade. We then discuss the tensions that exist within the literature and here refer to the more recent additions to the GVC case-study based literature. We then proceed to introduce our methodology in Section 2. Here we set out the research questions to be addressed, which to the best of our knowledge have not yet been tackled in the Southern African region. We introduce the data used to explore our research questions and methodology. We acknowledge some of the limitations with our methodology, which is a pragmatic one. These are essentially two-fold: first, they arise because of data limitations within vi the firm-level surveys and our inability to compare all indicators of interest across countries; and second, because of the need to match products listed within these surveys with trade data (both datasets use different codes). In Section 3 we provide an overview of trade flows on an intra- and extra-regional basis. We expected different products to be traded within intra- and extra-regional markets. Our identification of products traded mainly on an intra-regional basis substantiates the findings of others such as Stevens et al. (2013) in this regard. However, there does seem to be evidence of a type of production network in operation within the region, as different types of metal products are exported by South Africa to other SACU members, which also in turn export them back to South Africa. They do not however, feature within South Africa’s major extra -regional exports. Despite this, the links between these products and the automotive industry deserve further attention. The only evidence we have of vertically fragmented vertical trade within the region relates to the textiles and clothing industry. Otherwise, in terms of regional value chain development (with consumers and producers being located within the region) we have the most evidence related to more consumer orientated producers. To some extent this is surprising given the high share of foreign ownership within firms operating within the region. The extent to which these findings support the multi-chain upgrading hypothesis within the case-study based GVC is questionable. In Section 4 we use a logistic regression model to explore the degree of association between selected GVC participation and upgrading indicators on firms choice of marketing channel: domestic, intra- or extra-regional for Southern African firms. We identify a significant association for a limited number of firm-level indicators with firms subsequent choice of end market. Overall, the results of the analysis suggest that:  SACU firms are highly likely to experience major or severe customs and trade regulations compared to most other regions. As major and severe customs and trade regulatory barriers are experienced, the more likely it is that Southern African firms export on an intra- rather than an extra-regional basis.  Firms with more experienced managers are less likely to export on an intra-regional basis and more likely to export to extra-regional markets. We cannot control, however, for the nationality of firm managers.  Firms exporting on an extra-regional basis are more likely to be foreign owned compared to intra-regional exporters. Hence, investment may be a substitute for accumulated knowledge and skills. Put another way, a change in firm ownership and a new type of relationship with lead firms may be necessary in order to access some types of GVC. The strength of this result overall is driven by the fact that all exporters in South Africa have a share of foreign ownership.  The more likely firms find an inadequately trained workforce as a major or severe barrier to their operations, the less likely they supply intra-regional markets. The reasons for this deserve further investigation.  Firms supplying intra- rather than extra-regional markets seem more likely to provide formal training to their workers, though this result is not significant.it is to some extent suggestive of particular constraints for local firms supplying regional markets in terms of recruiting and retaining skilled workers, despite being more likely to provide formal training. vii Table 1: Summary of Results: Odds Ratios Global SACU SACU Indicator SACU Enterprise AFR EAP EECA LAC SAR Intra Extra Survey Ownership 50.247** 29.599** 64.115** 3.389** 3.892** 5.818** 4.348** 2.820** 1.595* Dummy Managers Experience 0.986 0.978** 1.022** 1.010** 1.026** 1.019** 1.020** 1.002 1.014** (years) Firm age 1.014** 1.006 0.985 1.007** 1.002 1.000 1.019** 1.009** 0.993* (years) Customs, TradeReg 1.768* 1.917** 0.257** 1.207** 1.000 1.389** 1.747** 1.535** 0.986 Dummy LabourReg 1.152 1.477 1.328 1.070 0.641** 1.300* 1.349** 0.961 1.219 Dummy Workforce 0.721 0.401** 1.108 0.868** 1.390 0.679** 1.003 0.895 1.049 Dummy Formal Training 1.298 1.332 1.124 1.308** 2.959** 0.656** 1.537 2.083** 1.730** Dummy Note: ** denotes significant at 5 percent level; * denotes significance at the 10 percent level. There are differences in the goodness of fit indicators for the model as applied to each region and these are discussed in Section 4. We have excluded discussion of multicollinearity and residual analysis. Source: World Bank Country Enterprise Surveys and Global enterprise Surveys; UN Comtrade. In order to further substantiate these results we integrate our results for SACU firms within an African sample, and make use of the World Bank Global Enterprise Surveys. This means we are able to compare the results we have obtained for Southern Africa to other African firms. Finally, we compare our results for Southern Africa, and Africa to those of other regions including: East Asia Pacific, Latin America and Caribbean, South Asia, and Europe and Central Asia. These results are summarised in Table 1. The odds ratio presents us with a measure of association rather than causation. Although we are unable to identify the causal relationship, and direction of causality, we can say there is a significant association. We also note though, that this association could be spurious. This caveat applies to all of the logistic regression analysis presented in this paper. Despite this caveat it is fair to say that the results overall indicate a level of heterogeneity across firms included in the Enterprise Surveys according to the type of GVC they export to (domestic; export; intra-regional exports; extra-regional exports). Our identification of products traded mainly on an intra-regional basis substantiates the findings of others, such as Stevens et al. (2013). However, there does seem to be evidence of a type of production network in operation within the region, as different types of metal products are exported by South Africa to other SACU members, which also in turn export them back to South Africa. These products do not however, feature within South Africa’s major extr a-regional exports (within the same product heading). Despite this, the links between these products and the automotive industry deserve further attention. The only evidence we have of vertically fragmented vertical trade within the region relates to the textiles and clothing industry. Otherwise, in terms of regional value chain development (with consumers and producers being located within the region) we have the most evidence related to more consumer orientated products. To some extent this is surprising given the high share of foreign ownership within firms operating within the region. The extent to which these findings support or reject the multi-chain upgrading hypothesis within the case-study based GVC is questionable at the current time. The multi-chain hypothesis posits that domestic firms may have more opportunities to launch their own manufactured and branded products within their home market or in neighbouring markets, with similar levels of development. The types of lead firms driving intra-regional value chains as conventionally understood (with consumers and producers within the region) and their relationships with foreign investors within SACU could be further explored. viii Some generic as well as specific policy recommendations arise from this analysis. Comparing the results for SACU to other African exporters, as well as other regions included in the Enterprise Survey, we can see that exporters are more likely to experience high or severe barriers to exporting in terms of customs and trade regulations, and this increases the odds of exporting more on an intra- compared to extra-regional basis. Non-tariff barriers applied in the region are already known to affect intra-regional exporters to a greater extent than intra-regional exporters (Keane et al., 2010). It is important to note that our quantitative analysis is underpinned by the assumption that firms have a choice in relation to their export market, and that this applies equally to intra- or extra- regional markets. In practice however, we know this choice is not equally weighted and the more qualitative and case-study based GVC literature emphasises such aspects as power, rents and barriers to entry, in this regard. Understanding input:output processes at the firm-level and value added processes at this level could help to shed further light on the constraints of domestic firm’s integration with GVCs, and promote further dialogue with the lead (global) firms that drive these. We not explored the role of imports into production from extra-regional suppliers, although we know these sources have grown relative to African suppliers. This aspect of research deserves further attention. .Overall the exploratory analysis presented in this Working Paper could be improved in a number of ways. Although we have described the theory which suggests differences in firm-level upgrading trajectories depending on the end market, and provided some empirical evidence which substantiates this view, case-study analysis would help to substantiate further the findings presented in this paper. Panel data analysis for countries with the available data may be more revealing. The distinction between intra- and extra- regional exporters for other regions included in the Global Enterprise surveys could also be undertaken in the future. Finally, the robustness of results could be improved, with some variables excluded from analysis so as to sharpen the results. ix The new wave of GVC literature The emergence of the mainstream literature on GVCs is an acknowledgment of the increased importance of intra-firm trade in global trade, which is an unprecedented phenomenon. This trend is a result of the fragmentation of global production across countries, as they have become integrated within the global trading system through trade and finance linkages. The new trade literature, which revealed firm-level heterogeneity, has touched upon these issues, but the GVC approach to analysis provides a more receptive and less abstract means of conceptualisation for policy makers. The latest wave of GVC literature has been accompanied by new descriptive analyses, and datasets, which are introduced briefly in the following sub-sections. However, as we emphasise, there remain some major shortcomings regarding analysis of sub-Saharan Africa’s (SSA’s) participation in current GVCs. Other tensions are glossed over. These include regarding economic power, and governance - aspects emphasised by the 1990s wave of GVC literature as influencing firm-level upgrading opportunities. This literature which is based on case-study analysis rather than input:output tables question the emphasis on GVCs, compared to regional and domestic value chains. This is because of evidence that suggests upgrading, and learning, opportunities for domestic firms serving domestic and regional value chains (or multi-chains) may be more forthcoming than within GVCs, driven by lead firms. Trading within GVCs characterised by hierarchical governance structures, may offer domestic firms rapid product and process upgrading, but the ability to obtain more functions within the value chain over time may be more limited. Despite this, even if upgrading processes are limited to nodes of production, entering into GVCs may help to expand formal employment opportunities for low-skilled workers. This applies to countries which would otherwise be unable to expand domestic production. These aspects have not been explored to the best of our knowledge in the Southern Africa region. In this Section, we first briefly review the new wave of GVC literature, and highlight the new insights it offers, as well as shortcomings. We discuss generally why GVCs matter for regions, countries and firms. We then discuss some of the recent findings for the Southern African region and members of the South African Customs Union (SACU). We then elaborate upon the posited links between GVC development and regional value chains in terms of upgrading opportunities and challenges. Here we make the link to the firm-level literature. Finally, we discuss why governance structures matter, in terms of influencing upgrading trajectories for firms trading within different types of value chains: global, regional, or national. Hence, we discuss how and why better supporting regional value chain development may assist integration processes with GVCs. ODI Report 1 1.1 Why GVCs matter for regions, countries and firms1 The term GVC has become wax lyrical in recent years. This includes within different schools of thought and methodological perspectives. Although originating from a sociological perspective and rooted within world systems theory, it has subsequently become used by business analysts so as to better understand processes of value creation and sources of comparative advantage and how to maximise these (e.g. Porter 1985). Managing linkages within value chains effectively can increase profit margins. Support services such as infrastructure, human resources, technology development and procurement became just as important as core business activities. The trade community became alerted to just how integrated supply-chains are because of the effects of the global financial crisis of 2008 and the synchronised global trade collapse which occurred. Since then GVCs have been used to reinvigorate the trade policy debate (e.g. countries need to import before they export) and have focused attention on behind the borders issues. For example, Hoekman and Jackson (2013) argue that trade policy makers need to “think supply chains” and recognise the integrated and interconnected nature of global trade patterns. This literature tends to focus on intermediate goods trade, or vertically fragmented trade. More qualitative development economists became interested in GVCs from around the early 1990s as developing countries became more deeply integrated into the global economy. This literature was concerned about notions of economic power and asymmetries in trading relations. It is derived from Global Commodity Chain analysis and world systems theory, which emphasises how economic relationships are constructed over time, rather than emerging spontaneously (Wallerstein, 1974). The term GVC was used in the 1990s as analysis shifted towards understanding how value creation can be influenced by local processes (Gereffi, 1999). Different GVC governance structures were identified based on where economic power resides within the value chain. Within this literature, there are some commonalities towards the analysis of value chains across sectors including manufactured, agricultural and extractive-industries (Kaplinsky and Morris, 2001). Despite these differences, all perspectives recognise that global trade patterns have never been so interconnected. Moreover, that trading success within increasingly integrated global markets means entering and upgrading within GVCs. The ability of countries to do so depends on many of the policy measures applied at their borders, as noted by the traditional trade literature, as well as considerations related to institutions and geography. However, success also requires consideration of new issues and beyond the border measures, such as the effective management of foreign direct investment (FDI). New Trade Opportunities As discussed by the WTO (2013) multilateral, regional and unilateral trade liberalisation has greatly increased market access and, together with sharply falling transportation and communication costs, this has facilitated the emergence of value chains. Production that once was primarily located close to sources of major suppliers of inputs (or near consumers in final markets) is now increasingly carried out wherever the necessary skills and materials are available at competitive cost and quality. This fragmentation process has created new opportunities for developing countries to enter global markets as components or services suppliers, without having to build the entire value chain. By providing access to networks, global markets, 1 This section of the report is drawn from Keane (2014a). ODI Report 2 capital, knowledge and technology, integration in an existing value chain can provide a first step to economic development – a path that is often easier to travel than building a complete value chain (OECD, 2013). Shares in Intermediate Goods Trade Latest estimates suggest that around 80% of all trade takes place within the international production networks of transnational corporations (TNCs), around one-third of which is intra-firm trade – that which occurs within the ownership structure of a single firm, or TNC (UNCTAD, 2013). Intermediate goods trade accounts for 60% of global trade (ibid). Developing countries now account for around half of global trade flows: developing economies accounted for only 34% of world merchandise exports in 1980, but by 2011 their share had risen to 47%, or nearly half of the total (WTO, 2013). At the current time around 85 percent of trade in value added – or intermediate goods trade – takes place in and around three hubs and the three regional blocks of East Asia, Europe and North America (AfDB et al., 2014; 127; Baldwin, 2012). SSA’s level of participation in this new wave of trade is worryingly low. It has increased which suggests that the continent as a whole is engaging more with this new wave of intermediate goods trade. But, the share of African suppliers’ in the continents imports has been declining compared to imports sourced from outside Africa; these flows have grown twice as fast as exports (OECD, 2014; 74). In sum, these results suggest particular changes for African producers in entering into GVCs and upgrading within them. Table 2: Share of Trade in Value Added by Region, 1995 and 2011 (percentage) Region 1995 2011 Europe 57.5 50.9 East Asia 14.4 16.2 North America 13.1 11.8 Southeast Asia 6.0 6.8 Latin America 3.2 4.2 Middle East 2.0 3.0 Africa 1.4 2.2 Russia and Central Asia 0.9 2.0 South Asia 0.7 1.7 Oceania 0.9 1.3 Source: AfDB et al. (2014), calculated based on UNCTAD-EORA GVC database The Economics of GVCs The cost considerations that underpin the fragmentation process of production have been analysed by trade economists. For example, Baldwin and Venables (2013) explore the technological characteristics of products and economic geography considerations; they emphasise the presence of centripetal forces that bind some activities together – a process that differs across products, and depends on the co- location of certain activities.2 They make a distinction is made between ‘snakes’ production processes compared to ‘spider’ processes. In the snake production process a physical entity follows a linear process with value added at each stage, e.g. cotton production to yarn to fabric to t-shirts manufacture. On the other hand, in the case of ‘spiders’, production process may be many limbed. This is where parts from different sources come together in one place for assembly; this may not be the final destination, as any part of a spider might be attached to any part of a snake (e.g. buttons). 2 See Keane (2014b) for discussion of recent new slants old phrases with economic geography connotations. ODI Report 3 They recognise that there are recognised tensions between the comparative costs that create the incentive to unbundle compared to the co-location or agglomeration forces that may bind some parts of a process together. The fragmentation of stages of the production process is therefore determined by opposing forces of international cost differences and the benefits of co-location of related stages. The end result will depend on the technological relationships between stages of production. Given recent trends on the proportions of trade controlled by TNCs one assume that increasingly complex and technologically sophisticated products are being produced in fragmented chains, but that control by lead firms remains high (through FDI), either because domestic capabilities in recipient countries remain low, or because the benefits of co-location and agglomeration forces remain weak relative to costs (since governing value chains costs lead firms). Both of these aspects - cost differences and benefits of co-location – could therefore assist in explaining Africa’s current participation in GVCs characterised by intermediate goods trade. 1.2 GVCs in SSA and SACU Despite the headline figures for SSA, there are some reasons for optimism, and this is particularly the case in the Southern African region. Intra-African trade in value added manufactured goods - or intermediate goods trade - has grown, and more rapidly than flows destined for the rest of the world (AfDB et al., 2014). The Southern African region is a key driver of these trends as we discuss below. South Africa as a driver South Africa is seemingly playing the role of headquarter economy in the region: its share of intermediates from other economies in the region increased nine-fold between 1995 and 2011 (from USD78million to USD686million); South African intermediates embedded in the exports of other economies increased five-fold over the same period (from USD675million to USD3,487million).3 According to the OECD (2012) and their estimates of GVC participation, South Africa performs better than China (Figure 1). Their GVC participation index measures the relative importance of imports and intermediate exports in total exports, or the extent to which countries are involved in vertically fragmented production4 Figure 1: Participation in global value chains Source: OECD/WTO TiVA database (2012) 3 See AfDB et al. (2014). 4 This is approximated by the sum of (i) the value of imported inputs in the overall exports of a country and (ii) the percentage of exported goods and services used as imported inputs to produce other countries’ exports. The shares added describe the participation of a country in GVCs, both as a user of foreign inputs and as a supplier of intermediate goods and services used in other countries’ exports. ODI Report 4 Regional Dynamism It is estimated by the AfDB et al. (2014) that Southern Africa, accounts for 40% of Africa’s total GVC participation, one third of which is through backward integration: the share of foreign value added embedded in a country’s exports. Although overall, Europe and Asia are found to be the main sources of foreign value added embedded in African exports, this trend is bucked in Southern Africa. The reasons for the concentration of GVC participation in Southern Africa compared to other regions in Africa are not explored further. The data created by the AfDB et al. (2014) in their backward integration matrix suggests that other regional partners contribute a range of between 0.01 - 0.1 percent of imported value added into South Africa’s exports. On the other hand, South Africa provides more than 10 percent of intermediate goods in Botswana, Namibia, Swaziland, Zambia and Zimbabwe. Although useful information is provided by the AfDB et al. (2014) the types of products, sectors, and lead firms driving these value chains is not discussed in detail. Moreover, the link to trade policy tends to be rather prescriptive, e.g. simply corresponding to the current mainstream discourse of the need to import before exporting, or describing how the penalties for infant industry protection are higher now than in the past (because countries specialise in tasks rather than complete value chains). Although it is acknowledged that African participation in GVCs is limited to lower value activities within GVCs, no particularly innovative policies are identified to overcome this. 1.3 Upgrading: Why it matters, and link to GVC Governance The increased phenomenon of intra-firm trade essentially translates into more trade within hierarchical GVC governance structures (Keane, 2012). Although the hierarchical type of GVC governance is posited to result in rapid product and process upgrading, the route to functional upgrading – moving towards a higher position within the value chain – is posited to become more difficult. This is essentially, because this route towards upgrading is blocked by lead firms who control more lucrative nodes of production (Tijaja, 2012; Kamau 2009; Navas- Aleman, 2011). A typology of a GVC governance structures articulated by Gereffi et al. (2005) is briefly described in Box 1 below. A hierarchical GVC governance structure as defined in this framework (and so widely referred to within the GVC literature), means that lead firms exert a high degree of control on their suppliers; this includes through ownership and backward vertical integration. It is important to note that this typology of governance was developed on the basis of case-study analysis, and has not been subject to much more empirical scrutiny. Although hierarchical structures of governance have received much attention within North-South trade analyses, it is fair to say that other types of governance structures have received less attention; this includes more relational types of governance which may exist within South-South value chains. ODI Report 5 Box 1: Governance of Global Value Chains The types of GVC governance identified by Gereffi et al. (2005) are derived from country case-study analyses. The typology is useful as it highlights the degree of explicit coordination within a GVC, which is a function of the producers’ capabilities, the complexity of the transactions and the extent to which they can be codified. Market governance is typical where transactions are relatively simple, information on product specifications is easily transmitted, and suppliers can make products with minimal input from buyers. Trade takes place through arms-length exchanges which require little or no formal cooperation, for example, carried out in auction houses or other spot markets, or other over-the-counter transactions. Modular governance occurs when complex transactions are relatively easy to codify. Suppliers in modular chains make products to a customer’s specifications. Information technology and standards for exchanging information are both key to the functioning of modular governance, which means greater control over transactions relative to the market-based governance. Relational governance occurs when buyers and sellers rely on complex information that is not easily transmitted or learned. This results in frequent interactions and knowledge sharing between parties. Lead firms specify what is needed, and exert some level of control over suppliers but relational linkages take time to build, so the costs and difficulties required to switch to a new partner tend to be high. This type of governance is typically associated with contract manufacturing, but offshore contractors may source inputs. Captive governance is a feature of chains where small suppliers are dependent on one or a few buyers that often wield a great deal of power. Such networks feature a high degree of monitoring and control by the lead firm. The power asymmetry in captive networks forces suppliers to link to their buyer under conditions set by, and often specific to, that particular buyer. Some cases of contract farming or types of commodity trade can exhibit this type of governance, given low supplier competence and complex transactions. Hierarchical governance usually occurs when product specifications cannot be codified, products are complex, or highly competent suppliers cannot be found. This type of governance is typically associated with industries where all stages of production are carried out ‘in house’ as production is offshored rather than outsourced. Source: Adapted from Gereffi et al. (2005). Firm Ownership Structures The more recent additions to the case-study GVC literature now draw particular attention to the concept of social embeddedness, and why firm ownership matters (Kaplinsky and Wamae, 2010; Staritz and Morris, 2013; Morris and Staritz, 2014). This perspective on the motivations of investors and their embeddedness, the destination end markets and subsequent influence on upgrading trajectories is also shared by Goto et al. (2009). Generally, the emerging literature on the motivations of investors relates these to potential upgrading trajectories, either at the node of production, or across. For example, depending on the relationship developed with investors and lead firms, there may be greater learning opportunities and possibilities of technology transfer. But this depends on investors’ motivations and the extent to which these have progressed from merely efficiency or resource- seeking but towards relationship building. These concerns are clearly similar to the literature on FDI spillovers. Farole and Winkler (2014) discuss these issues within the context of GVCs. They argue that FDI’s spillover potential—the productivity gain resulting from the diffusion of knowledge and technology from foreign investors to local firms and workers— is one of the most valuable inputs to long-run growth and development. Their conceptual framework is built around an understanding of the mediating factors that shape the nature and extent of spillovers, specifically: the spillover potential of foreign investors (particularly in the context of investments within GVCs); the absorptive capacity of local agents (firms and workers); and how these two factors interact within a specific host country institutional environment. Clearly, these are similar areas to those included in the GVC governance framework developed by Gereffi et al., (2005). ODI Report 6 Box 2: Market and Investor Diversification With reference to the development of the textiles and clothing industry in Cambodia, Goto et al. (2009) conclude that exporting to the Japanese market rather than the US market might facilitate more efficient technology transfer for firms in the Cambodian garment industry. This is because the approach of investors and nature of relationships between firms is qualitatively different. Figure 1a presents the market segmentation for Cambodian garments developed by Goto et al. (2009). Figure 1a: Market Segmentation Classification for Cambodian Garments Source: Goto et al. (2009) It is notable, however, that the regional trade policy dimension is not explicitly referred to in this study. The reasons for this are not clear, although we know that Cambodia is a late industrialiser to the Asia Pacific region, where the ‘flying geese’ model of recycling comparative advantage and integration of countries within regional production networks - referred to in detail by IDE-JETRO (2011) - has driven economic development. Making the link to firm-level performance matters, because although within the qualitative GVC analysis, the ownership of firms, investors’ motivations and end markets are receiving greater attention, it is fair to say the link to actual firm-level data remains weak and based mostly on a descriptive rather than a quantitative analysis. Where quantitative analysis of employment within GVCs has been undertaken, it has resulted in some very broad-brushed statements on the superior performance of GVC integrated - or FDI driven firms - and employment outcomes (e.g. Shepard and Stone, 2013). Firm-level Performance Generally, firm-level studies continue to explore the existence or not of learning by doing effects; it is fair to say that the trade and growth literature has come full circle because of the results. The recent debate on trade and growth at the level of the firm not only posits that the most productive firms within an industry export, but also that such firms are likely to have been the most productive firms within an industry before exporting, i.e. exporters self-select into markets. This is in part due to the high sunk costs for firms to enter markets.5 Moreover, some studies find that productivity typically increases to a greater extent before firms export, as opposed to after firms enter export markets (Clerides et al., 1998; Bernard and Jensen 2004; Greenaway and Kneller, 2007). New trade theory is distinguished from old new trade theory by its focus on the firm as opposed to the industry; it focuses more on internal economies of scale to the firm as opposed to external economies to the industry which may result from the clustering of firms (for example, as in new economic geography models). That is, it is more in line with models of monopolistic competition that incorporate heterogeneous firms and the theory of the multinational enterprise. However, as opposed to considering all firms as homogenous exporters it introduces firm heterogeneity: some firms export, others don’t, some firms are larger and more 5 Because of external economies of scale: established patterns of specialization may persist even though they run counter to comparative advantage (e.g. Swiss watch producers compared to Thai). ODI Report 7 productive than others. For example, Antras and Helpman (2004) explore the implications of firm heterogeneity for the boundaries of the firm and strategies for outsourcing and insourcing of activities. Markusen and Venables (2000) develop a monopolistic competition model of trade in which the presence of trade costs changes the pattern of trade, creates incentives for factor mobility which may lead to an agglomeration of activity in a single country and may therefore lead to multinational firms. Further to the recent contributions by Melitz (2003), Helpman et al. (2004) and Bernard, Eaton et al. (2003), which reconsiders how firms participate in international markets, the set of firm characteristics analysed has been extended to include: size, age, human capital, capital-intensity and ownership. However, it is noted by Greenaway and Kneller (2007:42-44) that: “Although the evidence unambiguously points to the crucial role played by sunk costs, little research has focused on what these are and how agglomeration, exchange rates and policy changes affect them…other evidence is much less conclusive, such as that relating to learning by exporting.” Some of the few GVC studies that integrate firm-level analyses and make reference to the effect of governance structures on firm-level productivity include Pietrobelli and Saliola (2008). These authors engage with the issue of causality and the learning by exporting hypothesis; they find that the governance of value chains affects the productivity of suppliers in domestic value chains to a greater extent than for firms supplying MNEs or exporters. They note that this result may be explained by the different nature of information and knowledge being exchanged, and by the larger gaps in knowledge and capabilities between the domestic leader and its suppliers.6 Essentially, the case-study GVC literature describes how domestic firms must collaborate with lead firms in order to access and export to global markets. Given recent trends on the extent of intra-firm trade and that which takes place within the networks of MNEs, this means engaging with hierarchical GVCs which are driven by a few lead firms. Strategic alliances can be sought to achieve mutually beneficial outcomes. Policies can influence firm-level upgrading trajectories. Developing Upgrading Indicators Given the theoretical and empirical developments within the more macro and case- study GVC literature, and integration of firm-level analyses, there have been increased efforts to untangle just exactly what it means to upgrade within GVCs.7 Because of these efforts have been made to assign quantitative indicators to the qualitative GVC upgrading typology (Kaplinsky and Morris, 2001; Humphrey and Schmitz, 2004) which include the following:  Product upgrading: performing certain tasks better and improving quality of products.  Process upgrading: improving efficiency and quantities through improving production processes.  Functional upgrading: acquiring skills that enable movement towards another node of production, e.g. from production to marketing.  Inter-sectoral upgrading: using skills acquired to move into another sector, e.g. moving from clothing production to motorcycle manufacture. 6 As found by Pietrobelli and Saliola (2008) firms that service the domestic, rather than export market, can also benefit considerably from engagement with lead firms operating within GVCs. 7 However, some challenges remain since movement from one functional position can only be known through detailed case-study and data analysis over time. ODI Report 8 In addition to this, Bernhardt and Milberg (2011) distinguish between economic and social upgrading. Within the framework they develop, economic and social upgrading are defined as follows:  Economic upgrading: trade performance, as indicated by export unit values and market shares.  Social upgrading: employment and wage growth. The economic upgrading indicators are essentially adapted from Kaplinsky and Santos-Paulino (2005). Moreover, the analytical approach is essentially limited to linking economic and social upgrading to a particular node of production as opposed to viewing the movement of labour (and investors) across and into new functions.8 Although the results from this research endeavour are insightful, the ability to monitor these processes over time remains challenging. Movement from one functional position can only be known through detailed case-study analysis, and data analysis over time, and this upgrading strategy is excluded from their framework. Multi-Chain Upgrading There is an emerging literature on “multi-chain” upgrading. This relates to the greater learning opportunities available to firms serving multiple markets. In particular, domestic firms may have more opportunities to launch their own manufactured and branded products on domestic or neighbouring markets, with similar levels of development. This literature draws on the experience of producers in the textiles and clothing industry in Kenya (Kamau 2009) and furniture and footwear industry in Brazil (Navas-Aleman, 2011). Participation in multiple value chains provides the possibility of “leveraging competencies”: different value chains create different possibilities for learning, and what is learned in one value chain can be applied in others (Lee and Chen, 2000). A focus on domestic markets leads manufacturing firms to broaden the scope of their activities (i.e. functional upgrading) into design, marketing, and branding. This may be because they have a better understanding of home markets than foreign markets, or it may be because domestic customers are not as powerful or concentrated as their counterparts in global value chains (Brandt and Thun, 2010). The variance of governance types in end markets is related to buyer as well as consumer demands. Hence, some aspects of the Gereffi et al. (2005) framework (which is based on Northern markets) may be relevant, notably relations with lead firms. However, one limitation of the ‘multi-chain’ literature is that it obscures how and why FDI can be a substitute for domestic capabilities. For countries that with extremely limited productive capabilities, attracting FDI and entering into GVCs at a particular stage of production in order to begin capital accumulation and assimilation processes, remains an important new trade opportunity. Overall though, this literature concurs with our hypothesis that value chain governance types vary according to end markets and this affects firm-level productivity, and hence subsequent upgrading processes. 1.4 The link between regional and global value chains Given the opportunities as well challenges which are acknowledged within the literature regarding engaging with GVCs, particularly more tightly controlled and hierarchical types, our hypothesis is that regional markets and value chains may be more conducive to some types of upgrading processes by domestic firms, than 8 See: http://www.capturingthegains.org/ ODI Report 9 tightly controlled global markets. Hence, we expect there to be differences in firm- level characteristics according to whether firms export products on an intra-region compared to extra-regional basis. These aspects have been discussed briefly in Stevens et al., (2013) with regards to the development of African value chains. However, with regards to Africa more generally, the case-study based GVC literature tends to either:  remain focused on the effect of the Asian drivers on African producers (Fu et al. 2012; Kaplinsky, 2010), or  refer to new demand patterns driven by the BRICS, including South Africa, with reference to sectors that are already well covered by the literature such as textiles and clothing (Morris et al. 2014; Staritz and Morris, 2013). Where reference is made to building local and regional value chains in Africa and within a GVC context, the implied logic seems to be so as to prepare firms to subsequently export globally. For example, UNCTAD (2013) notes that “local and regional value chains have vital roles to play in broadening the manufacturing base of African economies, expanding productive capacity and boosting intra-African trade” (UNCTAD, 2013:92). Furthermore that: “Regional value chains present opportunities for improving productivity…both for domestic firms with export potential and those that produce goods predominantly demanded at the national and regional levels. For domestic firms…”regional value chains give them the opportunity to upgrade and achieve international competitiveness, thereby making it easier to connect with GVCs” (UNCTAD, 2013:92). Hence, from this perspective, regional markets are understood to provide the environment through which producers may learn and develop their productive capabilities and capacity before integrating with GVCs. Demand factors as well as competitiveness considerations are emphasised by UNCTAD (2013). Despite this recommendation, it must be recognised that some countries (in particular the LDCs) may have such limited domestic capabilities that attracting regional FDI and situating this within an FDI-led development strategy might in practice be the most viable strategy, at initial stages of economic development. The source of FDI and investor motivations would therefore need to be carefully scrutinised, and integration processes carefully managed. Getting data on investment in Southern Africa, however, can be even more challenging that obtaining accurate trade data. RI and GVCs: Theory and Practice As developed by Viner (1950) in his model of a customs union, regional integration can create trade diversion and enhance the supply of less competitive regional firms that are now protected by a common external tariff at the expense of competitive international supply. Much stronger statements are made regarding the development of more producer-driven9 GVCs and regional integration processes by others. For example, Humphrey and Memedovic (2003) find the emergence of regional production systems in the automotive industry itself resulted in regional integration. They make reference to these processes as driving regional integration process with reference to the Triad regions: North America; the European Union 9 Those where the economic power resides in control at the node of production. ODI Report 10 (EU), and Japan.10 These processes can be seen to continue with a self-reinforcing dynamic, as most trade in value added remains within these regions. It is well known that the history of European regional integration was driven by a single industry: coal and steel. In comparison, open regionalism in Asia was very much driven by the trade within the regional economic communities of intermediate goods, without the use of regional apparatus such as a free trade agreement or customs union. Instead, the regional integration process was driven by the private sector. This included through the use of more bottom-up initiatives that first defined, and then focused on the removal of barriers for priority products that mattered for regional production networks and GVC integration. In the case of Latin America and in particular Mercosur, Ciravegna (2003) finds that both regional and global value chain integration seem to have provided positive stimuli for product and process upgrading. As predicted by the GVC governance framework, inserting a production unit in a global chain introduces pressures to upgrade. However, the typical negative implications that may arise when inserting an economic unit into a GVC did not manifest themselves. The hierarchical structure of GVC governance was mediated, and the headquarter firms did not absorb more links of the value chain. Instead, in view of the Mercosur treaties, a concerted effort was undertaken to ensure that the Brazilian plant acquired certain links of the chain, such as product development, and some decisional autonomy. This study therefore highlights the importance of regional integration and interventions that can be designed at this level. As we are acutely aware the motivations for deeper regional integration in Southern Africa are substantially different to other regions such as Asia, or the EU: SADC, with a membership of 15 countries, was originally conceived as a bulwark against apartheid South Africa and, by extension, the Southern African Customs Union (SACU). SACU itself is not a Common Market, nor an Economic Union. Four out of the five members of SACU operate within a Common Monetary Area comprising Lesotho, Namibia, South Africa and Swaziland, but excluding Botswana, which has its own currency. This is not a common monetary area in the conventional sense: the currencies are pegged at par against the South African Rand and not convertible outside their territories; i.e. basically all these countries use the Rand. The SADC Trade Protocol was established in 1996, and since 2000 members have started to implement their commitments; a free trade agreement was launched in 2008. All SACU members are also members of SADC. The new evidence on GVC participation in the region as elaborated upon by AfDB et al. (2014) suggests that South Africa is a key driver of regional value chains, both as a supplier of inputs and as a destination market for exports. But we don’t know much about the nature of integration of other economies within SACU or types of products traded. Nor do we know much about the relative merits in terms of firm-level upgrading trajectories of supplying intra- compared to extra-regional markets. We intend to shed further light on these aspects in the following sections of this report. 1.5 Concluding Remarks The more macro descriptive GVC analysis, and the case-study GVC literature, introduced in the preceding sections has so far failed to critically engage with the intra-regional trade debate, particularly in Southern Africa. This brief literature 10 See: http://www.unido.org/fileadmin/user_media/Publications/Pub_free/Global_automotive_industry_value_chain.pdf ODI Report 11 review has demonstrated why we should expect differences in firm-level performance indicators according to the value chain served: intra- or extra-regional. We expect different products to be traded within these markets; these will be associated with different GVC governance structures and hence, subsequent upgrading opportunities. The research methodology deployed to explore these aspects next is introduced in the following Section. ODI Report 12 Methodology: Firm- level evidence on value chain participation The World Bank Enterprise Survey data contain a number of indicators of interest to this study. There are some limitations, however, in terms of the availability of all indicators of interest across SACU members. Despite this, these surveys are the most representative across the region to date, and are also publicly available. We therefore make use of them in order to better link GVC to firm-level analyses, though we remain somewhat cautious in terms of interpretation of our results in some instances. Despite lots of new input:output tables to describe GVC participation between countries, the link to firms and their characteristics remain rather weak (in spite of recent trends in intra-firm trade). Although our primary interest is to research potential and actual differences between intra- and extra-regional value chains, we also compare firms included in the Global Enterprise Survey in general. This is so as to draw out differences in the GVC participation related indicators we identify as being of interest, and as introduced in the following sub-Sections 2.1 Use of Enterprise Surveys The indicators used for the descriptive and quantitative analysis were identified based on their relevance to the research questions, as well as data availability. Because of data limitations we have not been able to include in the analysis additional indicators such as use of website to connect with clients. The surveys record firm-level information for those operating in the manufacturing sector, as well as retail and services.11 For the purposes of this analysis we are particularly interested in the manufacturing sector for the reasons discussed in the previous section. Other sectors such as retail are inevitably involved in trade as importers, but not necessarily as exporters. Although the tourism sector, which is included in the services data, trades it doesn’t trade in specific products and services exported by firms are not recorded. Hence it will not be possible in this instance to link exports to specific ISIC codes, which is necessary so as to distinguish between extra and intra-regional exporters and subsequently so as to explore performance differences between firms. Data are available for at least one period for all SACU countries, and two periods for Botswana (however, we make use only of the most recent year):  Botswana, 2010;  Lesotho, 2009; 11 Sector: Food, Garments, Other Manufacturing, Retail, Other Services, Rest of [firm] Universe. ODI Report 13  Namibia, 2006;  South Africa, 2007; and  Swaziland, 2006. The total number of firms surveyed for each of the countries of interest is summarised in Table 2 below. Table 3: Firms included in Enterprise Survey Country Total Firms Surveyed Data Utilised Botswana 268 268 Lesotho 151 151 Namibia 106 106 1,057 firms in the Southern South Africa African sample;937 in the 1,057 Global Enterprise Survey 70 (Manufacturing); 307 (All: Swaziland 70 Global Enterprise Survey) 1652 Note: Survey data are available for two periods in the case of Botswana (2006 and 2010) though we make use of the nearest years’ data. There are some slight differences in the availability of data included in the country and global survey results for South Africa and Namibia. We use the country surveys in order to explore differences in firm-level characteristics for intra- and extra-regional exporters. We then use the Global Enterprise Survey to compare all regions. Although information on end markets recorded in the Enterprise Surveys, it is patchy. Firm ownership is an indicator of GVC participation because we expect foreign-owned firms to be more deeply integrated in GVCs (intra-firm trade). Firms’ uses of imports are also recorded. The enterprise surveys do contain some information on the nationality of firms’ major owners, which we make use of to the extent possible. The data therefore enables us to identify firms according to their market orientation and ownership structure as follows:  Exporter Type: Direct sales = 10% or more of sales are exported  Ownership Type: Domestic, 10% or more foreign ownership Data on specific descriptive as well as performance indicators are provided for firms differentiated by these characteristics across the broad categories collected by the World Bank Enterprise Surveys, which include:  Biggest Obstacles reported by firms;  Corruption;  Crime;  Finance;  Firm Characteristics;  Gender;  Informality;  Infrastructure;  Innovation and Technology; ODI Report 14  Performance;  Regulations and Taxes;  Trade; and  Workforce. Table A1 Appendix reviews the data included in the Enterprise surveys across SACU members and identifies those indicators which are of interest to this study. This is undertaken if they meet at least one of the following three conditions:  indicative of GVC participation: e.g. firm ownership (share of foreign equity) is a strong indicator of GVC participation.  related to the expected outcomes of GVC participation including regarding the achievement of social and economic upgrading: e.g. formal training provided by firms can improve workers skills.  related to a policy/institutional constraint that can directly affect GVC participation and connection to external markets (intra and extra- regional) and be addressed through regional integration efforts: e.g. days to clear imports through customs could limit GVC participation. Through reviewing the Enterprise survey data in this way we were able to identify specific indicators of interest. We then checked their availability across SACU countries; this process meant we were unable to use many of the selected indicators simply because of lack of availability. Table 4 summarises the indicators we were able to utilise. Missing variables in the case of the numeric indicators - Age, Size, and Manager’s experience - were coded as ‘1’. In the case of categorical variables any missing variables were coded as ‘0’. Table 4: Data Utilised and Availability Botswana Lesotho Namibia South Africa Swaziland Indicator No. n/a No. n/a No. n/a No. n/a No. n/a Ownership 268 0 151 0 106 0 1057 0 70 0 Age 266 2 150 1 106 0 1056 1 68 2 Size (no. 268 1 147 4 105 1 1047 10 70 0 employees) Licence - - - - - - 680 377 - - International 5 (don’t Quality 263 151 0 106 0 936 121 70 0 know) Certification Length of Relationship with - - - - - - 1055 2 - - Suppliers Managers 267 1 150 1 105 1 1055 2 70 0 experience Education level of - - - - - - 1057 0 - - manager Formal training 268 0 151 0 106 0 1056 0 70 0 (l10) Workforce education: 268 0 151 0 106 0 1057 0 70 0 obstacle? Labour market regulations: 268 0 151 0 106 0 1057 0 70 0 obstacle? Customs and trade regulations: 268 0 151 0 106 0 1057 0 70 0 obstacle? Note: - means not available. The purpose of the analysis carried out in this paper is intended to focus attention on the soft infrastructure of regional integration (policy and institutional development) as opposed to hard aspects such as the facilitation of access to electricity or transportation systems. We are interested in indicators beyond those normally reviewed as part of the “doing business” assessments. Nevertheless, we still checked the availability of indicators such as connection to websites and use of ODI Report 15 the internet to connect with customers. Because of limited data we have been unable to incorporate these indicators into the analysis. Table 4 presents the indicator we have had to exclude from our analysis. In relation to the use of direct imports, data are extremely patchy, and so we do not make use of them. The responses are a bit better regarding the use of inputs: foreign and domestic. However, there are still coding issues which mean we have to exclude these indicators from our analysis. This is also the case regarding information on wages. Because we cannot include analysis of wages in our analysis, we are essentially limited to analysis of social upgrading. It is for this reason, we also explore recent trends in unit values and market shares for the major products trade don intra- and extra-regional markets for SACU countries. Table 5: Data Not Utilised Botswana Lesotho Namibia South Africa Swaziland Indicator No. n/a No. n/a No. n/a No. n/a No. n/a Percent of firms identifying access 268 0 151 0 106 0 1057 0 70 0 to finance as a major constraint Proportion of working capital - - - - - - - - - - financed by supplier credit (%) - Percent of firms having their own website (do you have internet 268 0 0 0 0 0 1057 0 0 0 connection C23) Percent of firms using email to interact with clients/suppliers c24c 194 74 0 0 0 0 153 904 0 0 – do you use the internet to deliver services Real annual sales growth (%) - - - - - - - - - - - Annual employment growth (%) Annual labour productivity growth - - - - - - - - - - (%) - Days to obtain an import licence 20 228 25 81 23 83 57 1000 14 56 j11 Days to clear imports from 0 268 74 77 73 33 680 377 35 35 customs d14 Days to clear direct exports 22 246 36 115 27 79 921 136 26 44 through customs –d4 Proportion of foreign inputs that 87 181 151 0 106 0 800 257 70 0 are of foreign origin d12b Number of temporary workers l6 268 0 151 0 106 106 937 120 70 0 Proportion of unskilled workers l4b 87 181 0 0 105 - 680 377 70 0 Monthly compensation 0 0 268 151 0 106 2 1057 0 70 (production)* Note: - means not available’ * similar data availability for monthly compensation levels for professionals and managers. As has already been noted in Section 1, the term ’GVC’ is becoming almost synonymous with FDI. However, as we have argued, direct ownership of firms through FDI is only one type of GVC and would constitute those characterised by the deepest form of integration - intra-firm trade - and by a hierarchical style of governance. Firms may still participate in GVC trade, but not necessarily through the means of direct ownership and FDI. The first step in our analysis was therefore to prepare the SACU dataset organise firms according to whether or not they export, and then to explore differences in our indicators of interest. Questions are asked in some of the enterprise surveys regarding the origin of foreign investors, as well as their ethnicity. However, analysis becomes difficult in terms of comparing the results across SACU members because of differences in the exact questions asked across the enterprise surveys. Not all of the enterprise surveys asked questions on the ethnicity of owners (e.g. Botswana does, South Africa just asks if African or not, similarly in the case of Lesotho). Despite these limitations, the majority of investors in Lesotho are of other African origin, followed by Other Asian. In the case of Botswana, most investors are “African”, followed by European and Asian investors. ODI Report 16 A shortcoming of the questionnaire in the Enterprise Survey, as it currently stands, is that it does not ask questions related to the nationality of firm managers. As we will see later in the report this becomes important as the experience of managers is identified as a significant variable in the quantitative analysis undertaken. This result is interesting because within the new trade/new growth literature, firm manager’s play an important role in diffusing knowledge and creating knowledge spillovers. In brief, although we have expressed interest in many of the indicators included within the World Bank Enterprise surveys. But due to data limitations we are limited in our ability to make full use of all of these. We first undertook an exploratory and descriptive analysis of the data for firms in SACU and then compared these to the results of firms in other regions included in the Global Enterprise Survey. These results are discussed in Chapter 3 which follows. Table 6 below provides information on the availability of the indicators of interest for other regions. We dealt with missing data in the same way as we did for SACU countries (coding categorical variables as 0 and numerical data as 1). Table 6: Availability of Indicators in the Global Enterprise Survey Dom./ Export Manag Customs Age of Labour Workforce Formal Region Foreig Dumm ers and trade Size firm regs. Educ. Training n y Exp. regs. Africa exc. Southern Africa Total 13958 13958 13958 13958 13958 13958 13958 13958 13958 Missing 0 0 348 313 0 0 0 194 0 Africa inc. Southern Africa Total 15950 15950 15950 15950 15950 15950 15950 15950 15950 Missing 0 0 361 325 3 3 3 203 4817 East Asia Pacific Total 8644 8644 8644 8644 8644 8644 8644 8644 8644 Missing 25 202 327 153 1 1 0 18 1647 Eastern Europe Central Asia Total 15682 15682 15682 15682 15682 15682 15682 15682 15682 Missing 161 143 523 153 6 6 5 120 1867 Latin America and Caribbean Total 25587 25587 25587 25587 25587 25587 25587 25587 25587 Missing 1180 61 1066 259 8 17 20 82 4536 South Asia Total 4345 4345 4345 4345 4345 4345 4345 4345 4345 Missing 10 129 49 37 295 64 61 15 7 Note: Because we only made use of the most recent years survey data this means we excluded the following in Africa: Angola (2006), Botswana (2006), DRC (2006), DRC (2010), Kenya (2007), Mali (2007), Rwanda (2006), Tanzania (2006), Uganda (2006), Zambia (2007); in Latin America we excluded: Argentina (2006), Bolivia (2006), Chile (2006), Colombia (2006), Ecuador (2006), El Salvador (2006), Guatemala (2006), Honduras (2006), Mexico (2006), Nicaragua (2006), Panama (2006), Paraguay (2006), Peru (2006), Uruguay (2006), Venezuela (2006). In Eastern Europe Central Asia we excluded: Albania 2007, Armenia 2009, Azerbaijan 2009, Belarus 2008, Bosnia and Herzegovina 2009, Bulgaria 2007, Bulgaria 2009, Croatia 2007, Macedonia 2009, Georgia 2008, Kazakhstan 2009, Kosovo 2009, Kyrgyz Republic 2009, Latvia 2009, Lithuania 2009, Moldova 2009, Romania 2009, Russia 2009, Serbia 2009, Slovenia 2009, Tajikistan 2008, and Ukraine 2008. In South Asia we excluded: Afghanistan 2008, Bangladesh 2007, and Nepal 2009. For our quantitative analyses we were only able to use of the following:  Firm-level characteristics and GVC participation indicators: ownership, age, size  Social and Economic upgrading indicators: manager’s experience, workforce education  Policy and institutional barriers: labour market regulations, customs and trade regulations. ODI Report 17 Obviously the indicators included in these categories are debateable. For example, some may prefer size (number of employees) to be included as an indicator of social upgrading. However, we have organised the data in this way based on the results of our literature review, e.g. new trade theory posits that only the largest and most productive firm’s export to global markets. We then used a logit model to explore the influence of firm level indicators on the subsequent choice of marketing channel: domestic or export; we then further distinguished between intra-regional exports and extra-regional exports. Underpinning our model specification are the following major hypotheses, which are informed by Section 1 and the results presented in the above sub-sections, namely that that: H1: Alternative: there are significant differences in the selected indicators on firm’s decisions (odds ratio) or not to export to intra-regional markets compared to extra- regional. This is because exporting to intra-regional markets may offer domestic firms opportunities to achieve functional upgrading, whilst exporting to extra-regional market may result more in product and process upgrading at the node of production. H0: Null: there are no significant differences in the effect of the above indicators on firm’s decisions (odds ratio) or not to export to intra-regional markets compared to extra-regional. This is because similar firm-level capabilities are required to export to either market, and firms that export to both have achieved a type of multi-chain upgrading. We test this model across the whole sample of firms and the following marketing channels: o Domestic market; o Export market; o Intra-regional market; o Extra-regional market. We specify as a dichotomous outcome variable, coded as = 1 if firms produce and export, and use to denote the probability that firms export the said good. The probability of an alternative marketing channel being selected is therefore (1 – ) and hence the model is specified as: () = + 1 ownership + 2 age + 3 managers experience + 4 size + 5 customs, trade reg dummy + 6 labour reg dummy + 7workforce dummy + 8training dummy + We explore the influence of firm-level characteristics on the odds that firms select a particular marketing channel and present our results in the following sub-sections. This analysis is necessary so as to answer following sub-research questions and hypotheses. These have all been derived from the theory discussed in Chapter 1. GVC participation indicators: How do GVC indicators differ between firms that export products mostly to intra- or extra-regional markets? H1.1. Firms that export to extra-regional markets will be more likely foreign owned H0 1.1: There are no significant differences in the influence of firm ownership between intra- and extra-regional exporters ODI Report 18 H 1.2: Firms that export to extra-regional markets will be more likely larger (employees) H0 1.2: There are no significant differences in the influence of firm age between intra- and extra-regional exporters H 1.3: We posit that older firms may be more likely to supply intra-regional markets. The age of a firm is typically used as an indicator of learning by doing in the firm-level literature. This is because in order to access extra-regional markets alliances will lead firms must be sought, and FDI may be a substitute for accumulated knowledge and experience. H0 1.3: There are no significant differences in the influence of firm age between intra- and extra-regional exporters Social upgrading indicators: How do social upgrading indicators differ between firms that export products mostly to intra- or extra-regional markets? H1.4: Firms that export to extra-regional markets employ managers with more experience than those that export to intra-regional markets because accessing export requires accumulated tacit knowledge H0 1.4: There are no significant differences in manager’s experience between firms that supply intra- or extra-regional markets H1.5: Firms that export to extra-regional markets are more likely to find lack of workforce education a challenge compared to firms that export on an intra-regional basis H0 1.5: There are no significant differences in the extent to which firms consider an inadequately educated workforce as a major or severe constraint Policy and institutional barriers: How do policy and institutional barriers affect extra- and intra-regional exporters? H1.6: Firms that export to extra-regional markets are more likely to find customs and trade regulations as a major or severe constraint H0 1.6: There are no significant differences in the extent to which intra- or extra- regional exporters experience customs and trade regulations as major or severe barriers H1.7: Firms that export to extra-regional markets are more likely to find labour market regulations a major or severe constraint H0 1.7: There are no significant differences in the extent to which firms that export mainly on an intra- or extra-regional basis find labour market regulations a major or severe constraint We then tested this model specifying 1 as our marketing channel of interest. Finally, we integrated the results of the Global Enterprise Surveys in order to compare and contrast with the results of SACU countries which for ease we simply refer to as Southern Africa. Translating Product to Trade Codes The second step in our analysis was to organise firms according to whether they export on more of an intra or extra-regional basis. We use these terms to refer to intra-African and extra-African markets. This second step was hampered by two major (and one smaller) methodological difficulties. Whilst we were able to find a ODI Report 19 way around them, they do mean that the results must be treated with great caution as suggestive pointers to further field research rather than definitive statements. The first major methodological problem is that the goods produced by the survey respondents are classified according to the International Standard Industrial Classification (ISIC) whilst the trade data needed to observe direction of trade is recorded in the Harmonised System (HS).The second major difficulty is that the trade data show the direction of exports for all SACU firms whilst the survey covered only a limited sub-set of firms. This second problem was exacerbated by the first: the ISIC categories are much broader than those of the HS so the latter had to be aggregated into groups (Table 7) that will have included many goods that are not actually produced by the Survey respondents. The enterprise survey describes the economic activity that the firm is engaged in and its main product produced; it then provides a corresponding ISIC code. We have only made use of the main product reported in the enterprise survey (Product 1) and no others (including Product 2), though from a scan of the data most firms produce goods that fall within the same ISIC code though we have not analysed patterns in detail. Table 7 below provides an overview of the number of corresponding HS codes for each ISIC recorded for firm in our SACU sample. Table 7: First Product Produced by Firms that Export # of corresponding ISIC ISIC description of economic activity HS6 codes 1511 Production, processing and preserving of meat and meat products 80 1512 Processing and preserving of fish and fish products 70 1514 Manufacture of vegetable and animal oils and fats 52 1520 Manufacture of dairy products 24 1531 Manufacture of grain mill products 26 1533 Manufacture of prepared animal feeds 2 1541 Manufacture of bakery products 6 1543 Manufacture of cocoa, chocolate and sugar confectionery 12 1549 Manufacture of other food products n.e.c. 33 1554 Manufacture of soft drinks; production of mineral waters 4 1711 Preparation and spinning of textile fibres; weaving of textiles 358 1721 Manufacture of made-up textile articles, except apparel 59 1722 Manufacture of carpets and rugs 23 1729 Manufacture of other textiles n.e.c. 60 1810 Manufacture of wearing apparel, except fur apparel 232 1920 Manufacture of footwear 29 2022 Manufacture of builders' carpentry and joinery 6 2109 Manufacture of other articles of paper and paperboard 33 Publishing of books, brochures, musical books and other 2211 8 publications 2221 Printing 7 Manufacture of basic chemicals, except fertilizers and nitrogen 2411 489 compounds 2412 Manufacture of fertilizers and nitrogen compounds 32 2413 Manufacture of plastics in primary forms and of synthetic rubber 72 2421 Manufacture of pesticides and other agro-chemical products 5 Manufacture of paints, varnishes and similar coatings, printing ink 2422 20 and mastics Manufacture of soap and detergents, cleaning and polishing 2424 38 preparations, perfumes and toilet prepara 2429 Manufacture of other chemical products n.e.c. 116 2519 Manufacture of other rubber products 46 2520 Manufacture of plastics products 68 2693 Manufacture of structural non-refractory clay and ceramic products 9 2695 Manufacture of articles of concrete, cement and plaster 13 2710 Manufacture of basic iron and steel 205 2720 Manufacture of basic precious and non-ferrous metals 168 2811 Manufacture of structural metal products 8 2893 Manufacture of cutlery, hand tools and general hardware 72 2899 Manufacture of other fabricated metal products n.e.c. 122 2913 Manufacture of bearings, gears, gearing and driving elements 18 ODI Report 20 2930 Manufacture of domestic appliances n.e.c. 51 3110 Manufacture of electric motors, generators and transformers 33 3120 Manufacture of electricity distribution and control apparatus 19 3130 Manufacture of insulated wire and cable 9 3140 Manufacture of accumulators, primary cells and primary batteries 14 3190 Manufacture of other electrical equipment n.e.c. 41 Manufacture of television and radio transmitters and apparatus for 3220 11 line telephony and line telegraph 3410 Manufacture of motor vehicles 30 Manufacture of parts and accessories for motor vehicles and their 3430 17 engines 3610 Manufacture of furniture 23 3691 Manufacture of jewellery and related articles 20 3699 Other manufacturing n.e.c. 80 We piloted several approaches to minimising these problems and mismatch between HS/ISIC codes. Distinguishing per se between goods that are exported mainly inside and mainly outside Africa is not a problem. While there are many goods, of course, that are exported to both markets, a significant number of exports from SACU states are directed predominantly either to African or to non-African markets (see Stevens and Kennan 2013). Our initial approach began by identifying all HS6 codes exported by a SACU state (in the same year as that state’s Survey – so 2006 for Namibia and Swaziland, 2007 for South Africa, 2009 for Lesotho, and 2010 for Botswana) to a value of $1 million or more and identifying the ISIC codes to which they correspond. We then did a simple count of the ‘qualifying’ HS6 codes (i.e. those in which exports were valued at $1 million or more) within each ISIC code to determine how many were exported predominantly (50% or more of the value) to other African countries, and those which were exported predominantly outside Africa. This gave us three categories: 1. ISIC codes in which over 50% of exports in all of the qualifying HS6 heads were directed to Africa; 2. those in which over 50% of exports in all qualifying HS6 heads were directed outside Africa; 3. and those in which some of the qualifying HS6 heads were exported over 50% to Africa and others in which exports were mainly outside Africa. Focusing our analysis on just categories 1 and 2 would have been fairly robust. In many cases, the proportion of exports in an HS6 head destined for either Africa or the rest of the world was substantially greater than the 50% threshold. Hence an ISIC group composed entirely of such cases could reasonably have been portrayed as being made up of goods destined primarily for the African or the non-African market. Unfortunately, of the country/ISIC code combinations actually cited in the Surveys fell into category 1 and some into category 2. The remaining 34 fell into category 3; in other words they contained a mix of some HS6 heads that were exported primarily within Africa and some exported primarily outside, and so this approach had to be amended. The alternative from which the results in the next Section are derived omit the exclusion of HS6 heads that were exported to a value of less than $1 million. This brings into the net a large number of very small exports some of which, it would seem, are relatively important for the Survey respondents who reported themselves as exporters. Unlike the first approach, the second approach classified an ISIC group as being 'predominantly intra' or 'predominantly extra' according to the aggregate value of all ODI Report 21 covered HS6 exports rather than according to the number of HS6 codes that were exported predominantly within or outside Africa. We therefore calculated the total value of each SACU country’s exports of each ISIC group in Table X (by Commented [MO1]: Which table are you referring to? aggregating the value of each component HS6 head) and then split the group into two according to whether 50% or more of the value of these aggregate exports was within or outside Africa. This approach could have resulted in a misleading classification (for example if exports were split almost equally between intra- and extra-African markets such that an ISIC group in which 49.4% of the total was destined for Africa but the whole group was classified as ‘extra-African’). But this did not happen. For example, in the case of South Africa, for 19 of the 29 ISICs for which Intra or Extra was clearly established the difference between the shares going to Africa and the rest of the world was over 40%; for only 1 was it less than 8% with a couple at 8% and the next smallest difference being almost 15%. A third methodological problem was less serious because South Africa is the dominant source of the exports recorded. It is that data for intra-SACU trade is unreliable. An apparent South African export to Namibia, for example, might genuinely be destined for consumption in that state and hence, legitimately, be classified as intra-African. But it could equally be a product in transhipment to Walvis Bay/Namibia for onward shipment that has been erroneously classified (for a variety of reasons). In this case, since the ultimate destination is not known it is not possible to classify it as either intra- or extra-African. Because we had based almost all of analysis on the first approach to classifying firms according to the main destination of their major good produced, and used these results to analyse trade performance in Chapter 3, we cross-referenced the results of a 90% threshold using our old approach and the results with our new approach (which uses a 50% threshold) and found the same products to feature. Hence, we still make use of some of these results in Chapter 3 of this report, though we emphasise that these should be interpreted cautiously and in broad-brush terms. ODI Report 22 Extra and intra-regional trading patterns In this Section we link firms to specific products exported, and identify those which are traded mostly on an intra and extra-African basis in line with the methodology described previously. We first present analysis of the types of products traded most on an intra- and extra-regional basis. We then apply our 50% and 90% threshold to identify those products mostly traded on an intra or extra-African basis. 3.1 Products Produced by Exporting Firms Generally, the number of products that tend to be traded more on the extra-regional basis are less numerous and hence to some extent more concentrated. The main products exported mostly to extra-regional markets and produced by firms in the enterprise survey, as listed in Table 7, include: furniture, other manufacturing, processed fish and meat, apparel, car machinery. Table 8: Exporting Firms, and Major Products Traded on an Extra-Regional basis (50% threshold) Republic Row Grand Description Lesotho Namibia of South Labels Total Africa 3610 Manufacture of furniture 17 17 Manufacture of cutlery, hand tools and general 2893 hardware 16 16 1512 Processing and preserving of fish and fish products 8 8 1730/ 1810 Manufacture of wearing apparel, except fur apparel 6 6 1810 Manufacture of wearing apparel, except fur apparel 4 4 Manufacture of basic chemicals, except fertilizers 2411 and nitrogen compounds 4 4 Production, processing and preserving of meat and 1511 meat products 3 3 Manufacture of television and radio transmitters and 3220 apparatus for line telephony and line telegraph 2 2 2022 Manufacture of builders' carpentry and joinery 1 1 Manufacture of paints, varnishes and similar 2422 coatings, printing ink and mastics 1 1 2710 Manufacture of basic iron and steel 1 1 Manufacture of basic precious and non-ferrous 2720 metals 1 1 Manufacture of bearings, gears, gearing and driving 2913 elements 1 1 Manufacture of electric motors, generators and 3110 transformers 1 1 3691 Manufacture of jewellery and related articles 1 1 Grand Total 6 10 51 67 Note: calculated using the methodology set out in Section 2; Swaziland is excluded because no products meet the threshold. Source: World Bank Enterprise Surveys and UN COMTRADE database. ODI Report 23 Some of these products are relatively well covered within the existing GVC case- study literature (e.g. furniture). However, more recent performance in export markets has not been explored to the best of our knowledge. Nor has the extent to which these products may be accompanied by regional production networks been explored. Looking at the data another way, and applying a more stringent value, and destination market threshold, we can see that many of the same products listed in Table 8 above, also feature in Table 9 below. However, furniture no longer becomes the major extra-regional export from South Africa; similarly some of the motor vehicle components no longer feature within this threshold. Table 9: Exporting Firms, and Major Products Traded on an Extra-Regional basis (90% threshold) ISIC ISIC description Lesotho Namibia South Total Code Africa 1511 Production, processing and preserving of 3 3 meat and meat products 1512 Processing and preserving of fish and fish 8 8 products 1730/ Manufacture of knitted and crocheted fabrics 1810? and articles 6 6 Manufacture of wearing apparel, except fur apparel 1810 Manufacture of wearing apparel, except fur 4 4 apparel 2411 Manufacture of basic chemicals, except fertilizers and nitrogen compounds 4 4 2893 Manufacture of cutlery, hand tools and 16 16 general hardware 2929 Manufacture of other special purpose 1 1 machinery 3691 Manufacture of jewellery and related articles 1 1 Grand Total 6 9 29 43 Note: calculated using the methodology set out in Section 2; Swaziland is excluded because no products meet the threshold. Source: World Bank Enterprise Surveys and UN COMTRADE database. 3.2 Intra-Regional Products Table 9 presents the product/country combinations for our intra-regional exporters, using the 50% value threshold. We can see that Swaziland features as an intra- regional exporter of ISIC1810 whilst only South Africa did as an extra-regional exporter. Other products feature such as: manufacture of cocoa (ISIC1543), manufacture of plastics (ISIC2520), and manufacture of other fabricated metal products (ISIC2899). We expect this latter product to feature within an intra- regional production network related to the automotive industry. However, we expect the first few products mentioned to be intra-regional value chains in the more conventional sense (producers and consumers are located within the region). We then moved on to apply our more stringent threshold. From Table 10 we can see that all of the same products feature. ODI Report 24 Table 10: Intra-Regional Exports (50% threshold) Republic Row Grand Description Botswana Lesotho Namibia of South Swaziland Label Total Africa Manufacture of bakery 1541 products 18 18 Manufacture of wearing apparel, 1810 except fur apparel 2 11 13 Manufacture of structural metal 2811 products 12 12 Manufacture of 2520 plastics products 3 2 1 6 Manufacture of cocoa, chocolate and sugar 1543 confectionery 1 3 1 5 Manufacture of soap and detergents, cleaning and polishing preparations, 2424 perfumes and toilet 5 5 2221 Printing 4 4 Manufacture of other chemical products 2429 n.e.c. 4 4 Manufacture of dairy 1520 products 3 3 Manufacture of other 1549 food products n.e.c. 3 3 Manufacture of other fabricated metal 2899 products n.e.c. 1 1 1 3 Manufacture of grain 1531 mill products 2 2 1549 or Manufacture of other 2925? food products n.e.c. 2 2 Manufacture of made- up textile articles, 1721 except apparel 1 1 2 Manufacture of other 1729 textiles n.e.c. 2 2 Manufacture of other articles of paper and 2109 paperboard 1 1 2 Note: calculated using the methodology set out in Section 2. Source: World Bank Enterprise Surveys and UN COMTRADE database. Table 11: Intra-Regional Exports (90% threshold) ISIC ISIC description Botswana Namibia South Swaziland Grand code Africa total 1520 Manufacture of dairy products 3 3 1543 Manufacture of cocoa, chocolate and sugar confectionery 1 3 1 5 1810 Manufacture of wearing apparel, except fur apparel 2 11 13 2424 Manufacture of soap and detergents, cleaning and polishing preparations, perfumes and toilet preparations 5 5 2520 Manufacture of plastics products 3 2 5 2899 Manufacture of other fabricated metal products n.e.c. 1 1 2 Grand Total 10 8 14 17 49 Note: calculated using the methodology set out in Section 2. Source: World Bank Enterprise Surveys and UN COMTRADE database. ODI Report 25 Some of the same products discussed in Stevens et al. (2013) as mostly exported on an intra-regional basis also feature in our analysis (which uses the same approach as our analysis of extra-regional exports presented above). These products include:  ISIC 2520: manufacture of plastics;  ISIC 2424: manufacture of soap and detergents. These products were identified by Stevens et al. (2013) because there are several African exporters and because most of these products have relatively high regional shares of exports. In our analysis, we have been able to link these products to specific firms in Botswana, Namibia, and South Africa. Hence, we can confirm that these products are produced by firms included in the enterprise survey and are traded mostly on an intraregional basis, which substantiates the findings of Stevens (2013).12 It is likely that the products that we identify in the following sub-section also fall within one of these categories: being an intra-African value chain in the conventional understanding, or being part of a broader production network operating within a GVC. We therefore encounter similar methodological challenges to those discussed by Stevens (2013) which includes our inability to say:  If these goods are regional value chains in the sense that some African producers supply African consumers – as industrial, or final consumers.  or, instead these products feature as inputs to extra-regional exports. It seems there are some types of production networks in operation within the region. The following products are exported by South Africa to other SACU members, which also in turn export them back to South Africa:  Different types of manufactured metal products (relating to ISIC2899 and ISIC2811). The links between these products and the automotive industry deserves further attention.13 In comparison, the following products are exported to South Africa by other SACU members, and feature within South Africa’s extra-regional exports:  Apparel (ISIC1810) The following products are exported to South Africa by other SACU members, or vice versa but then not but then do not feature as subsequent exports according to our analysis, on an extra or intra basis. This suggests they may be more conventional regional value chains with consumers and producers based within the region (although they may utilise imported inputs):  Dairy products (IS520);  Manufacture of cocoa (ISIC1543);  Soap, detergents etc. (ISIC2424);  Manufacture of plastics products (ISIC2520). Summary of results This brief descriptive analysis has run through the main products and markets for the intra- and extra-regional exporters. The descriptive analysis presented thus far 12 This is the most up to date analysis of intra-regional trade flows to the best of our knowledge. 13 For example Barnes and Kaplinksy (2000) discuss developments in the automotive components industry. ODI Report 26 has shown us that different products tend to be traded more on an extra- compared to intra-regional basis. Moreover, those different types of value chain seem to be in operation. Some upgrading processes may be easier to achieve within intra-regional value chains, precisely because these markets are less dominated by a few lead firms. Hence, supplying processed (and branded?) chocolate may be possible for firms orientated to intra-regional markets, compared to extra-regional (where processed chocolate does not feature as a major export). Differences in the types of products traded within these alternate marketing channels may explain some of the differences we identify in terms of firm-level characteristics in the next Section. 3.3 Firm-Level Characteristics Here we explore differences in firm-level characteristics for firms that export predominantly on an intra-regional compared to extra-regional basis. We only present some of the main descriptive results in this sub-section, which serves as an introduction to the next Section which applies our quantitative analysis. 3.3.1 SACU Exporters and Non-Exporters Most of the firms included in the enterprise survey don’t export (Table 11). Hence, analysis of “only exporters” reduces our sample size considerably, and is likely to be reduced further as we split the exporter sample defined by their end market: intra- regional or extra-regional. Table 12: Summary of enterprise data for Southern Africa Country Sectors Total Ownership Type Domestic Foreign Non- Non- Foreign Dom. Exporter Exporter exporter exporter South All 1,057 137 920 920 0 13 124 Africa Namibia Manufacturing 106 28 78 70 8 17 11 Botswana All 268 129 139 126 13 111 18 Lesotho All 151 50 101 101 0 22 28 Swaziland Manufacturing 70 27 43 34 9 11 16 Total 1,652 371 1,281 1251 30 174 197 Source: World Bank Enterprise Surveys for each country, described in Section Two. In broad terms what can be seen from Table 11 is that overall the proportion of foreign owned firms is smallest for the largest market (and producer) in the region – South Africa, followed by Namibia. However, all exporters in South Africa have a share of foreign ownership. Swaziland has the highest proportion of firms that export, followed by Namibia and Lesotho. The results of comparative analysis across indicators of interest - GVC participation, social and economic upgrading, policy/institutional barriers - revealed some differences between firms that export compared to those that don’t within the SACU sample.14 In brief the following points may be noted:  GVC participation indicators o Firm age: no significant differences between exporting/non- exporting firms. o Firm size: no significant differences in the size of exporting/ non- exporting firms. o Exporters are generally more likely to have international quality certification, except in the case of Lesotho. 14 See Appendix 2. ODI Report 27 o Exporting firms in Namibia and South Africa are the least likely to operate under sub-contracting arrangements, other countries seem to be more likely. o Non-exporters tend to have the longest length of relationship with their suppliers.  Social/economic upgrading indicators o Exporters generally employ a higher share of skilled workers – those with tertiary education. o Exporting firms generally more likely to provide formal training, and pay more in terms of average wages for production workers (but not managers). o Exporting firms have managers with the most amount of experience, particularly in South Africa.  Policy and Institutional Barriers o Exporters overall, and particularly within the smaller SACU economies, are more likely to experience barriers across all of the indicators we analysed, including:  Workforce education.  Labour market regulations.  Customs and trade regulations. We compared these results to those of other regions included in the World Bank Global Enterprise Survey (Appendix 3). Further to this descriptive overview the following points may be noted:  GVC participation indicators: o Ownership: the proportion of exporting firms with foreign ownership is lower for other regions compared to SACU, e.g. proportionately lowest in South Asia and highest in Africa. o Firm age: There is a range of between 19 (East Asia Pacific) and 34 (Latin America and Caribbean) years of operation for other regions compared to a low of 15 in Lesotho and a high of 27 in South Africa. o Firm size (employees): The average size of firms is highest for Swaziland (412) and lowest for Botswana (116), which compares to a low in the Global Enterprise Surveys of Eastern and Central Europe (176) and a high of South Asia (517). o International quality certification: The proportion of SACU exporter firms with international quality certification is comparable to other regions.  Social upgrading indicators: o Managers experience: The results for SACU exporters range from 9 years (Swaziland) to 19 years (Botswana) compare to a low of 13 for Africa overall and a high of 21 years (Latin America and the Caribbean).  Policy and institutional barriers: o Workforce education: Range of 6.7 (East Asia Pacific) to 24 percent (Latin America and Caribbean) compared to a range of 12 percent (South Africa) to 41.9 percent (Botswana) reporting high or severe barriers in SACU. ODI Report 28 o Labour market regulations: Range of 8 percent (East Asia and Pacific) to 29 percent (Africa) compared to 6 percent (Botswana) to 32 percent (Namibia) reporting high or severe barriers in SACU. o Customs, trade regulations: Range of 8 percent (East Asia Pacific) to 36.7 percent (Latin America and Caribbean) compared to a range of 2 percent (South Africa) to 39 percent (Botswana) reporting high or severe barriers in SACU. 3.3.2 Comparison between Extra- and Intra-Regional Exporters Table 12 below shows the number of firms that we have been able to classify as either an intra- or extra-regional exporter, making use of our 50% threshold, discussed in Chapter 2. Table 13: Intra-regional/extra-regional exporters Extra- Missing data/exporting Intra-regional regional Country Total exporters less than £1mn exporters (90%) exporters (90%) South Africa 124 11 62 51 Namibia 19 2 9 10 Botswana 31 19 12 0 Lesotho 28 13 9 6 Swaziland 25 0 25 0 Total 227 45 117 67 Source: World Bank Enterprise Surveys and UN Comtrade Despite data limitations and given that we are unable to discuss results for 45 firms because of inability to categorise them as either producing goods traded on either an intra- or extra-regional basis, overall we can observe the following: GVC participation indicators o Firm age: there are no significant differences between intra and extra-regional exporters. o Firm size: there are no significant differences between intra- and extra-regional exporters. o International Quality cortication: No major differences between intra- and extra-regional exporters. Social Upgrading Indicators o Managers with more experience tend to be employed by extra- regional exporters. o But the share of tertiary education for managers seems to be higher for firms producing goods traded mainly on intra-regional markets compared to extra-regional ones (particularly in the case of South Africa). o There are no major differences between firms in terms of formal training offered to workers. Policy/Institutional Barriers o Customs and trade regulations appear to be greater barriers for firms trading goods on an intra-regional basis. o Lack of workforce education tends to be more of a barrier for firms exporting goods traded mainly on an extra-regional basis. o Labour market regulations tend to be more of an obstacle for extra compared to intra-regional exporters, except in the case of South Africa. Overall, the most interesting aspect of these results seem to reside in relation to the differences between intra- and extra-regional exporters in terms of social upgrading ODI Report 29 indicators, and the extent to which policy/institutional barriers are a major or severe constraint to trade. We explore the extent to which these variables exert a significant influence on the likelihood of exporting to either intra- or extra-regional markets in the next Section. ODI Report 30 Analysis of enterprise surveys 4.1 Methodology Our objective is to explore whether or not there is any link between the indicators we have identified15 - GVC participation, social/economic upgrading, and policy/institutional barriers - and the likelihood that a firm chooses one marketing channel compared to another. We undertake this analysis first for SACU exporters and non-exporters; we then compare SACU intra-regional and extra-regional exporters. Finally, we integrate the results of the Global Enterprise Surveys and compare Southern Africa to other regions. We use a logistic model since the results from the analysis are more intuitive for those interested in GVC analysis. We are able to interpret the odds ratio for each dependent variable as either increasing, or decreasing the likelihood that a particular marketing route is chosen. We specify as a dichotomous outcome variable, coded as = 1 if firms produce and export, and use to denote the probability that firms export the said good. The probability of an alternative marketing channel being selected is therefore (1 – ) and hence the model is specified as: () = + 1 ownership + 2 age + 3 managers experience + 4 size + 5 customs, trade reg dummy + 6 labour reg dummy + 7workforce dummy + 8training dummy + Because of data limitations, we limit the variables we examine, as specified above. Underpinning our model specification are the following major hypotheses:  H1: Alternative: there are significant differences in the effect of the above indicators on firm’s decisions (odds ratio) or not to export. o H0: Null: there are no significant differences in the effect of the above indicators on firm’s decisions (odds ratio) or not to export. The following sub-hypotheses and research questions may be noted. GVC participation indicators: How do GVC indicators differ between firms that export products mostly to intra- or extra-regional markets? H1.1. Firms that export to extra-regional markets will be more likely foreign owned. H0 1.1: There are no significant differences in the influence of firm ownership between intra- and extra-regional exporters. 15 Which is also informed by data availability across SACU members. ODI Report 31 H 1.2: Firms that export to extra-regional markets will be more likely larger (employees). H0 1.2: There are no significant differences in the influence of firm age between intra- and extra-regional exporters. H 1.3: We posit that older firms may be more likely to supply intra-regional markets. The age of a firm is typically used as an indicator of learning by doing in the firm-level literature. This is because in order to access extra-regional markets alliances will lead firms must be sought, and FDI may be a substitute for accumulated knowledge and experience. H0 1.3: There are no significant differences in the influence of firm age between intra- and extra-regional exporters. Social upgrading indicators: How do social upgrading indicators differ between firms that export products mostly to intra- or extra-regional markets? H1.4: Firms that export to extra-regional markets employ managers with more experience than those that export to intra-regional markets because accessing export requires accumulated tacit knowledge. H0 1.4: There are no significant differences in manager’s experience between firms that supply intra- or extra-regional markets. H1.5: Firms that export to extra-regional markets are more likely to find lack of workforce education a challenge compared to firms that export on an intra-regional basis. H0 1.5: There are no significant differences in the extent to which firms consider an inadequately educated workforce as a major or severe constraint. Policy and institutional barriers: How do policy and institutional barriers affect extra- and intra-regional exporters? H1.6: Firms that export to extra-regional markets are more likely to find customs and trade regulations as a major or severe constraint. H0 1.6: There are no significant differences in the extent to which intra- or extra- regional exporters experience customs and trade regulations as major or severe barriers. H1.7: Firms that export to extra-regional markets are more likely to find labour market regulations a major or severe constraint. H0 1.7: There are no significant differences in the extent to which firms that export mainly on an intra- or extra-regional basis find labour market regulations a major or severe constraint. H1.8: Firms that export to extra-regional markets are more likely to find an adequately trained workforce a major or severe constraint. H0 1.8: There are no significant differences in the extent to which firms that export mainly on an intra- or extra-regional basis find an adequately trained workforce a major or severe constraint. We test this model across the whole sample of firms, and as differentiated by the end market as follows: ODI Report 32 o Domestic market orientated; o Export orientated; o Intra-regional market orientated; and o Extra-regional market orientated. We explore the influence of firm-level characteristics on the odds that firms select a particular marketing channel and present our results in the following sub-sections. We then discuss the extent to which we should accept or reject the null and alternative hypotheses. The odds ratio Exp (B) can be interpreted in terms of the change in odds resulting from a unit change in the predictor(s). If the value is greater than 1 it indicates that as the predictor increases the odds of the outcome also increase (and the converse would apply if the value is less than 1).16 Essentially, it is a measure of association between the response variable (marketing channel) and our explanatory variables. That is, the odds ratio tells us the probability of a particular marketing channel being selected depending on the explanatory variables. The odds ratio presents us with a measure of association rather than causation. Although we are unable to identify the causal relationship, and direction of causality, we can say there is a significant association. We also note though, that this association could be spurious. This caveat applies to all of the logistic regression analysis presented in this paper. Despite this caveat it is fair to say that the results overall indicate a level of heterogeneity across firms included in the Enterprise Surveys according to the type of GVC they export to (domestic; export; intra-regional exports; extra-regional exports). It is important to note – as we have stressed in our methodology section - that a significant and positive impact of high or severe barriers to production from customs and trade regulations on exports may indicate reverse causation; exporting firms tend to be more sensitive to the effect of trade regulations rather than sensitive firms tend to export more. It is possible that firms which export attract more foreign investment, as opposed to foreign ownership led to exportation. Finally, firms that export more on an extra-regional basis may simply find demand for skills is higher in these markets. All of these caveats should be borne in mind regarding interpretation of our results. The Wald statistic which has a chi-square distribution tells use whether or not the coefficient for our predictors is significantly different from zero. In this sense it is analogous to the t-tests found in multiple regressions. In terms of goodness of fit, we refer to the Cox and Snell as well as Nagelkerke R square values. These statistics are referred to as a "pseudo-R" statistics. The former takes the sample size into account and can be interpreted as. It cannot however, reach a maximum of one. Hence, we also refer to the Nagelkerke R square values which adjust the Cox and Snell measure so that 1 can be achieved.17 4.2 Domestic and Export Orientated SACU Firms The results for specifying specify as a dichotomous outcome variable, coded as = 1 if SACU firms only supply the domestic market are presented in Table 13. These suggest that older firms are less likely to supply domestic markets, similarly those that are foreign owned. Firms that select the domestic market are also less likely to encounter major or severe customs and trade regulations. Although the result of 16 The odds ratio is obtained by exponentiating the estimate of B. 17 We also examine residuals using casewise listing, as well as obtain classification plots, though we do not present these results. Collinearity diagnostics were also undertaken through a review of variance inflation factors. ODI Report 33 firm size is significant, its influence in terms of the odds ratio is ambiguous: it neither increases nor decreases the odds of firms selecting the domestic market. Table 14: Orientated to Domestic Market B S.E. Wald Sig. Odds ratio Exp(B) Ownership Dummy -3.917 0.221 313.795 0.000** 0.020 Managers Experience (years) 0.14 0.009 2.424 0.119 1.014 Firm age (years) -0.14 0.006 4.846 0.028** 0.987 Customs, TradeReg Dummy -5.69 0.315 3.271 0.071* 0.566 LabourReg Dummy -0.141 0.302 0.219 0.639 0.868 Workforce Dummy 0.327 0.243 1.802 0.179 1.386 Size (no. of employees) 0.000 0.000 5.659 0.17** 1.000 FormalTrainingDummy -0.261 0.192 1.848 0.174 0.770 Constant 4.066 0.262 241.451 0.000 58.308 Note: N= 1,652. The labour reg, customs/trade reg, and workforce reg dummy variables take a value of 1 if firms report these policy and institutional variables to have either a severe or major (negative) effect on production. Goodness of fit indicators: Cox and Snell 0.281; Nagelkerke R Square 0.511. **= significant at 5 percent level ; * = significant at 10 percent level. Source: World Bank Enterprise Surveys and UN Comtrade. The results for specifying specify as a dichotomous outcome variable, coded as = 1 if SACU firms export are presented in Table 14. We can see that the alternative hypothesis must be accepted in the case of the influence of the following GVC participation indicators: ownership; Firm age; and Firm size (no. of employees); and Policy and Institutional Barriers: The customs/trade/regulation dummy. Foreign ownership of a firm exerts the strongest degree of association with a considerable increase in the odds (50 times) that a firm will export if it is foreign owned. Because the odds ratio increases so much as the firm ownership predictor variable does, it highlights the limited strength of association for the other predictor variables. This result is driven by the fact that all exporters in South Africa have a share of foreign ownership. However, because the Wald statistic is positive for the other predictor variables this means we must accept the related alternative hypotheses in these instances, although we acknowledge that odds ratio increases to a much lesser extent for these compared to firm ownership. Firms that export also tend to be larger, although the effect of firm size on the likelihood of exporting is ambiguous. Table 15: Export Orientated Odds ratio B S.E. Wald Sig. Exp(B) Ownership Dummy 3.917 .221 313.787 .000** 50.247 Managers Experience (years) -.014 .009 2.426 .119 .986 Firm age (years) .014 .006 4.849 .028** 1.014 Customs, TradeReg Dummy .570 .315 3.274 .070* 1.768 LabourReg Dummy .142 .302 .220 .639 1.152 Workforce Dummy -.327 .243 1.801 .180 .721 Size (no. of employees) .000 .000 5.668 .017** 1.000 FormalTrainingDummy .261 .192 1.848 .174 1.298 Constant -4.066 .262 241.458 .000** .017 Note: N= 1,652. The labour reg, customs/trade reg, and workforce reg dummy variables take a value of 1 if firms report these policy and institutional variables to have either a severe or major (negative) effect on production. Goodness of fit indicators: Cox and Snell 0.280; Nagelkerke R Square 0.509. **= significant at 5 percent level ; * = significant at 10 percent level. Source: World Bank Enterprise Surveys and UN Comtrade. ODI Report 34 4.3 Intra- and Extra-Regional SACU Exporters We then moved onto look at the potential differences between firms that export goods traded mainly on an intra- or extra-regional basis. In order to do this we first specified as a dichotomous outcome variable, coded as = 1 if SACU firms export predominantly on an intra-regional basis. The results are presented in Table 15. As before, we make use of the full SACU sample in order to perform this analysis, and include those firms which we have not been able to identify as either mainly supplying intra- or extra regional markets (discussed earlier). This means that the independent dichotomous variable is only coded 1 only for those firms that produce products we definitely know are mainly supplied on an intra-regional basis (using our 50% threshold). We use the same major hypotheses as specified above, namely that:  Alternative: there are significant differences in the effect of our selected indicators on firm’s decisions (odds ratio) or not to export to intra-regional (African) markets.  Null: there are no significant differences in the effect of our selected indicators on firm’s decisions (odds ratio) or not to export to intra- regional (African) markets. Table 16: Results of Logistic Regression for Intra-Regional Exporters B S.E. Wald Sig. Odds ratio Exp(B) Ownership Dummy 3.388 0.276 150.264 0.000** 29.599 Managers Experience (years) -0.023 0.011 4.241 0.039** 0.978 Firm age (years) 0.006 0.007 0.650 0.420 1.006 Customs, TradeReg Dummy 0.651 0.346 3.533 0.060** 1.917 LabourReg Dummy 0.390 0.338 1.330 0.249 1.477 Workforce Dummy -0.913 0.319 8.206 0.004** 0.401 Size (no. of employees) 0.000 0.000 0.124 0.725 1.000 FormalTrainingDummy 0.287 0.223 1.650 0.199 1.332 Constant -4.261 0.309 190.130 0.000** 0.014 Note: N= 1,652. The labour reg, customs/trade reg, and workforce reg dummy variables take a value of 1 if firms report these policy and institutional variables to have either a severe or major (negative) effect on production. Goodness of fit indicators: Cox and Snell 0.140; Nagelkerke R Square 0.349. **= significant at 5 percent level ; * = significant at 10 percent level. Source: World Bank Enterprise Surveys and UN Comtrade. The results presented in Table 15 for our intra-regional exporters, suggest that we must accept the alternative hypothesis relating to the following indicators: ownership; social upgrading; manager’s experience; and the following policy/institutional barriers: customs/trade regulations, and workforce regulations. These variables exert a significant influence (10 and 5 percent level, respectively) on their association with the likelihood that a firm exports on an intra-regional basis. These results are easier to make sense of when we compare them to the results for extra-regional exporters. In Table 16 we specify as a dichotomous outcome variable, coded as = 1 if SACU firms export predominantly on an extra- regional basis. ODI Report 35 Table 17: Results of Logistic Regression for Extra-Regional Exporters Odds ratio B S.E. Wald Sig. Exp(B) Ownership Dummy 4.161 0.529 61.848 0.000** 64.115 Managers Experience (years) 0.022 0.013 2.981 0.084* 1.022 Firm age (years) -0.015 0.011 1.958 0.162 0.985 Customs, TradeReg Dummy -1.359 0.640 4.513 0.034** 0.257 LabourReg Dummy 0.284 0.411 0.478 0.489 1.328 Workforce Dummy 0.103 0.338 0.092 0.761 1.108 Size (no. of employees) 0.000 0.000 4.184 0.041** 1.000 FormalTrainingDummy 0.117 0.285 0.169 0.681 1.124 Constant -5.875 0.572 105.373 0.000** 0.003 Note: N= 1,652. The labour reg, customs/trade reg, and workforce reg dummy variables take a value of 1 if firms report these policy and institutional variables to have either a severe or major (negative) effect on production. Goodness of fit indicators: Cox and Snell 0.105; Nagelkerke R Square 0.365. **= significant at 5 percent level ; * = significant at 10 percent level. Source: World Bank Enterprise Surveys and UN Comtrade. Taken together these results suggest we must accept the alternative hypothesis in the case of the following:  Ownership: Foreign owned firms are much more likely to export on an extra- rather than intra-regional basis;  Managers experience: the odds ratio suggests firms that employ more experienced managers are more likely to export on an extra-regional basis than intra-;  Firm age: older firms tend to be slightly less likely to export on an extra-regional basis (and hence more likely to export on an intra- regional basis) though this result is only significant at the 16 percent level;  Customs and trade regulations dummy: firms that rank these as major and severe barriers to production are much less likely to export on an extra-regional basis, and more likely to export on an intra-regional basis;  Firm size (number of employees): though this variable is significant, its effect on the odds ratio is ambiguous. We expect firms that trade predominantly on intra and extra-regional basis to exhibit differences related to the indicators we use as proxies for GVC participation, social upgrading, and policy/institutional barriers. We have revealed such differences through our descriptive as well as quantitative analysis. This analysis could be improved though in a number of ways. It is standard practice within the literature to introduce interaction effects in order to explore differences between population sub-groups within an overall sample. A challenge in doing so at this stage of the analysis however, is that we have only coded firms as either intra- or extra- regional exporters within the overall SACU exporter sample. That is to say, if firms export, they also export on an intra- (or extra- regional) basis. This means we are unable to introduce interaction terms and proceed further at this stage in the analysis. However, we can improve on this in the following sub-section where we pool the SACU sample with all other African firms included in the ODI Report 36 Global Enterprise Surveys in order to increase the sample size and explanatory power.18 4.4 SACU and Other African Exporters Within the African exporter’s sample, it is possible to introduce an interaction effect for SACU firms and compare these exporters to other African exporters. Although this means we are unable to shed any further light on the differences between intra- and extra-regional exporters within the SACU sample, we can in terms of how our selected indicators influence SACU exporters compared to other African exporters. In order to compare estimated coefficients between two different groups, such as SACU and other African it is standard to run regression for a pooled sample and include interaction terms of each regressor with the SACU dummy, e.g. ownership*SACU dummy, in addition to regressors themselves and the SACU dummy. If an interaction term is statistically significant, it indicates that the coefficient differs between SACU and non-SACU. We present our results for this next step in the analysis in Table 17 below. Table 18: SACU and African Exporters B S.E. Wald Sig. Odds ratio Exp(B) SACU Dummy -0.765 0.275 7.717 0.005** 0.466 Ownership Dummy 1.359 0.075 332.163 0.000** 3.892 SACU*Ownership 2.526 0.231 119.110 0.000** 12.503 Managers Experience 0.025 0.007 13.932 0.000** 1.026 SACU*Managers Experience -0.036 0.011 10.523 0.001** 0.964 Firm Age 0.002 0.003 0.652 0.419 1.002 SACU*Firm Age 0.005 0.007 0.586 0.444 1.005 Customs, TradeReg Dummy 0.000 0.190 0.000 0.998 1.000 SACU*Customs, TradeReg Dummy 0.552 0.367 2.252 0.133 1.736 LabourReg Dummy -0.444 0.163 7.435 0.006** 0.641 SACU*LabourReg Dummy 0.579 0.342 2.860 0.091* 1.784 Workforce Dummy 0.329 0.205 2.588 0.108 1.390 SACU*Workforce Dummy -0.655 0.318 4.245 0.039* 0.520 Size (no. of employees) -0.001 0.001 1.470 0.255 0.999 SACU*Size 0.001 0.001 3.626 0.057* 1.001 FormalTrainingDummy -0.791 0.138 32.984 0.000** 0.453 SACU*FormalTraining 1.085 0.236 21.152 0.000** 2.959 Constant -3.193 0.099 1039.968 0.000 0.041 Note: N= 15611. The labour reg, customs/trade reg, and workforce reg dummy variables take a value of 1 if firms report these policy and institutional variables to have either a severe or major (negative) effect on production. Goodness of fit indicators: Cox and Snell 0.061; Nagelkerke R Square 0.149. **= significant at 5 percent level; * significant at 20 percent level. Source: World Bank Enterprise Surveys for Southern African Countries and Global Enterprise Surveys for other African countries and UN Comtrade. 18 It should be noted that the sample we use for SACU is comprised of the most recent year’s data for each country. In most cases, these data correspond to Global Enterprise Surveys, except in the case of South Africa and Swaziland. Hence, we use our SACU sample and combine it with all African countries (except SACU) included in the Global Enterprise Surveys. In the next Section, we simply make use of all of the Global Enterprise Survey data. ODI Report 37 These results and interpretation of the specified interaction effects suggest there are significant differences in the impact (or association) of the following indicators for SACU firms compared to other African firms:19  GVC Participation Indicators o Ownership: There is a much stronger association between this variable and the likelihood of exporting for SACU firms than other African. o Size: Although significant for SACU firms the increase in odds for exporting firms is marginal.  Policy and Institutional Barriers o Labour Regulations: SACU firms that experience major or severe barriers are more likely to export, whereas other African firms are less likely to export. o Workforce Regulations: SACU firms which experience these as major or sever barriers to production are less likely to export compared to other African firms. o Formal Training: SACU firms which provide formal training are much more likely to export compared to other African firms. 4.5 Comparison of SACU with Global Enterprise Surveys The final step in our analysis was to compare the results of the logic regression for our SACU and African sample, to that of other regions. We do this initially for the full sample of firms included in the Global Enterprise Surveys: Africa, East Asia Pacific, Eastern Europe and Central Asia, South Asia, and Latin America and the Caribbean. Table 16 presents the results of specifying as a dichotomous outcome variable, coded as = 1 if firms across the Global Enterprise Surveys export. Table 19: Results of Logistic Regression for Global Enterprise Surveys Indicator B S.E. Wald Sig. Exp(B) Ownership Dummy 1.221 0.032 1478.564 0.000** 3.389 Managers Experience (years) 0.010 0.001 64.925 0.000** 1.010 Firm age (years) 0.007 0.001 86.433 0.000** 1.007 Customs, TradeReg Dummy 0.188 0.039 23.313 0.000** 1.207 LabourReg Dummy 0.067 0.042 2.539 0.111 1.070 Workforce Dummy -0.141 0.035 15.802 0.000** 0.868 Size (no. of employees) 0.000 0.000 227.059 0.000** 1.000 FormalTrainingDummy 0.268 0.027 95.575 0.000** 1.308 Constant -2.671 0.028 8924.471 0.000** 0.069 Note: N = 59278. The following regions are included: Africa; East Asia Pacific, Eastern Europe Central Asia, Latin America and Caribbean, South Asia. Goodness of fit indicators: Cox and Snell 0.041; Nagelkerke R Square 0.078. **= significant at 5 percent level; * significant at 20 percent level Source: World Bank Global Enterprise Surveys and UN Comtrade. We can see that the odds ratio for the ownership variable for all firms in the Global Enterprise Surveys is similar to that of the African sample (Table 18) though much lower compared to our SACU exporter sample (Table 14). An increase in firm managers’ experience slightly increases the odds of exporting, and this is similar to 19The result for manager’s experience is approximately the same for African firms and SACU firms and hence we do not discuss it. ODI Report 38 the result for the Africa sample (Table 17). As a firm’s age increases, so too do the odds of exporting, though the increase in odds overall for firms is marginal. The labour regulations dummy is not significant, which is a key difference compared to the African sample (Table 17). However, the customs and trade regulations dummy is significant and this result is comparable to SACU exporters (Table 14). Firm size is ambiguous. Firms that provide formal training are more likely to export and this is similar to our SACU sample (Table 14). Summary of Comparative Analysis We summarise the results for SACU exporters, compared to Africa exporters as well as those across the regions included in the World Bank Enterprise Surveys in Table 19. We simply summarise the odds ratios across our indicators of interest and denotes whether or not they exert a significant influence. Finally, the results of our analysis of the individual regions included in the Global Enterprise Surveys are presented in Tables 20-23 overleaf. In each case, we specified as a dichotomous outcome variable, coded as = 1 if firms export. Table 20: Summary of Results: Odds Ratios Global Indicator SACU AFR EAP EECA LAC SAR Survey Ownership 2.820** 1.595* 50.247** 3.389** 3.892** 5.818** 4.348** Dummy Managers 1.002 1.014** Experience 0.986 1.010** 1.026** 1.019** 1.020** (years) Firm age 1.009** 0.993* 1.014** 1.007** 1.002 1.000 1.019** (years) Customs, 1.535** 0.986 TradeReg 1.768* 1.207** 1.000 1.389** 1.747** Dummy LabourReg 0.961 1.219 1.152 1.070 0.641** 1.300* 1.349** Dummy Workforce 0.895 1.049 0.721 0.868** 1.390 0.679** 1.003 Dummy Size (no. of 1.000** 1.002** 1.000** 1.00** 0.999 1.001** 1.000 employees) FormalTraining 2.083** 1.730** 1.298 1.308** 2.959** 0.656** 1.537 Dummy Note: ** denotes significant at 5 percent level; * denotes significance at the 10 percent level. There are differences in the goodness of fit indicators for the model as applied to each region. See Appendix 3. Source: World Bank Enterprise Surveys and UN Comtrade. East Asia and Pacific Surprising, an additional year of operating for firms in Asia Pacific (Table 20) does not increase the odds of exporting. Firm managers experience matters though: as manager’s experience increases so too do the odds of exporting. Firm size has an ambiguous effect on the odds of exporting. All of our policy/institutional barrier indicators are significant, though only the lack of workforce skills variable is associated with a reduction in the odds that firms export. ODI Report 39 Table 21: East Asia and Pacific Indicator B S.E. Wald Sig. Exp(B) Ownership Dummy 1.761 0.078 505.480 0.000** 5.818 Managers Experience (years) 0.018 0.004 26.910 0.000** 1.019 Firm age (years) 0.000 0.003 0.005 0.945 1.000 Customs, TradeReg Dummy 0.329 0.126 6.763 0.009** 1.389 LabourReg Dummy 0.262 0.155 2.878 0.090* 1.300 Workforce Dummy -0.388 0.128 9.203 0.002** 0.679 Size (no. of employees) 0.001 0.000 93.403 0.000** 1.001 FormalTrainingDummy -0.422 0.071 34.852 0.000** 0.656 Constant -2.373 0.080 871.437 0.000** 0.093 Note: N = 8554. Goodness of fit indicators: Cox and Snell 0.094 ; Nagelkerke R Square 0.168. Source: World Bank Enterprise Surveys and UN Comtrade. Eastern Europe and Central Asia Exporters in Eastern Europe (Table 21) are not only likely to find customs and trade regulations a major or severe barrier to exporting, but are also likely to find labour regulations problematic. The workforce dummy is not significant in this instance. Nevertheless, foreign ownership, number of years of managers’ experience and firm age are all significantly associated with an increase in the odds of firms exporting. Table 22: Eastern Europe Central Asia Indicator B S.E. Wald Sig. Exp(B) Ownership Dummy 1.470 0.072 419.299 0.000** 4.348 Managers Experience (years) 0.020 0.002 67.482 0.000** 1.020 Firm age (years) 0.019 0.002 123.510 0.000** 1.019 Customs, TradeReg Dummy 0.558 0.076 53.975 0.000** 1.747 LabourReg Dummy 0.300 0.087 11.849 0.001** 1.349 Workforce Dummy 0.003 0.062 0.002 0.966 1.003 Size (no. of employees) 0.000 0.000 14.827 0.000 1.000 FormalTrainingDummy 0.430 0.051 71.806 0.000 1.537 Constant -3.026 0.060 2558.363 0.000 0.048 Note: N = 15682. Goodness of fit indicators: Cox and Snell 0.55; Nagelkerke R Square 0.103 Source: World Bank Enterprise Surveys and UN Comtrade. Latin America and the Caribbean Foreign ownership increases the odds of exporting for Latin America and the Caribbean (Table 22), though to a lesser degree than East Asia Pacific and Eastern Europe and Central Asia. Firm age matters. There is a significant association between firms that experience severe or major barriers to production because of customs and trade regulations and their likelihood of exporting, though the odds ratio is lower than in the case of Eastern and Central Europe analysed above. Firms that provide formal training are more than two times likely to export. ODI Report 40 Table 23: Latin America and Caribbean Indicator B S.E. Wald Sig. Exp(B) Ownership Dummy 1.037 0.060 303.402 0.000** 2.820 Managers Experience (years) 0.002 0.002 1.074 0.300 1.002 Firm age (years) 0.009 0.001 60.818 0.000** 1.009 Customs, TradeReg Dummy 0.428 0.058 55.038 0.000** 1.535 LabourReg Dummy -0.040 0.058 0.462 0.497 0.961 Workforce Dummy -0.111 0.052 4.510 0.034 0.895 Size (no. of employees) 0.000 0.000 27.839 0.000** 1.000 FormalTrainingDummy 0.734 0.052 199.131 0.000** 2.083 Constant -2.688 0.066 1644.801 0.000 0.068 Note: N = 14657. Goodness of fit indicators: Cox and Snell 0.060; Nagelkerke R Square 0.104. Source: World Bank Enterprise Surveys and UN Comtrade. South Asia Contrary to expectations an additional years of operation for firms in South Asia (Table 23) would seem to decrease the likelihood that these firms export. The reasons for this deserve more attention. Interesting, this result is somewhat similar to that of our SACU sample. Firm managers experience matters in terms of increasing the odds that a firm exports, but the result is not as strong as for other regions including Southern Africa. Foreign ownership of a firm matters as this too increases the odds a firm exports, and this result (and strength) is similar to that of Southern African exporters. Finally, none of the policy and institutional barriers included in the logit regression are significant. Table 24: South Asia Indicator B S.E. Wald Sig. Exp(B) Ownership Dummy 0.469 0.2489 3.591 0.058* 1.595 Managers Experience (years) 0.014 0.005 7.531 0.006** 1.014 Firm age (years) -0.007 0.004 3.376 0.066* 0.993 Customs, TradeReg Dummy -0.014 0.156 0.008 0.927 0.986 LabourReg Dummy 0.198 0.184 1.160 0.282 1.219 Workforce Dummy 0.048 0.138 0.119 0.730 1.049 Size (no. of employees) 0.002 0.000 134.267 0.000** 1.002 FormalTrainingDummy 0.548 0.112 24.084 0.000** 1.730 Constant -2.599 0.124 440.083 0.000 0.074 Note: N = 4345. Goodness of fit indicators: Cox and Snell 0.091; Nagelkerke R Square 0.170 Source: World Bank Enterprise Surveys and UN Comtrade. ODI Report 41 Conclusions This paper has developed a research methodology to explore firm-level performance across different types of value chains: those destined for extra- regional markets, compared to those destined for regional markets. It has utilised the firm-level information included in the World Bank Enterprise Surveys for SACU members in order to do so. We have compared the results obtained for SACU to other regions in the Global Enterprise Surveys. Using this methodology, which admittedly uses some strong assumptions, we have revealed some differences across the groups of firms defined by the export market, which arguably warrant further attention. Our identification of products traded mainly on an intra-regional basis substantiates the findings of others, such as Stevens et al. (2013). However, there does seem to be evidence of a type of production network in operation within the region, as different types of metal products are exported by South Africa to other SACU members, which also in turn export them back to South Africa. These products do not however, feature within South Africa’s major extra-regional exports (within the same product heading). Despite this, the links between these products and the automotive industry deserve further attention. The only evidence we have of vertically fragmented vertical trade within the region relates to the textiles and clothing industry. Otherwise, in terms of regional value chain development (with consumers and producers being located within the region) we have the most evidence related to more consumer orientated products. To some extent this is surprising given the high share of foreign ownership within firms operating within the region. The extent to which these findings support or reject the multi-chain upgrading hypothesis within the case-study based GVC is questionable at the current time. The multi-chain hypothesis posts that domestic firms may have more opportunities to launch their own manufactured and branded products within their home market or in neighbouring markets, with similar levels of development. The types of lead firms driving intra-regional value chains as conventionally understood (with consumers and producers within the region) and their relationships with foreign investors within SACU could be further explored. We have revealed differences in firm-level characteristics according to their end market, and hence identified GVC related firm-level heterogeneity. Our approach to analysis has revealed a number of significant measures of association, though we are unable to identify the direction of causality (and association in some cases could be spurious). This caveat applies to all of the logistic regression analysis presented in this paper. Despite this caveat it is fair to say that the results overall indicate a level of heterogeneity across firms included in the Enterprise Surveys according to the type of GVC they export to (domestic; export; intra-regional exports; extra- regional exports). Some generic as well as specific policy recommendations arise from this analysis. There are differences between intra- and extra-regional exporters in terms of their ODI Report 42 firm-level characteristics, and the products traded between these markets differ substantially. We have identified some specific product/country combinations which may warrant further investigation. Comparing the results for SACU to other African exporters, as well as other regions included in the Enterprise Survey, we can see that exporters are more likely to experience high or severe barriers to exporting in terms of customs and trade regulations, and this increases the odds of exporting more on an intra- compared to extra-regional basis. Non-tariff barriers applied in the region are already known to affect intra-regional exporters to a greater extent than intra-regional exporters (Keane et al., 2010). The quantitative analysis carried out is underpinned by the assumption that firms have a choice in relation to their export market, and that this applies equally to intra- or extra-regional markets. In practice however, we know this choice is not equally weighted and the more qualitative and case-study based GVC literature emphasises such aspects as power, rents and barriers to entry, in this regard. Understanding input:output processes at the firm-level and value added processes at this level could help to shed further light on the constraints of domestic firm’s integration with GVCs, and promote further dialogue with the lead (global) firms that drive these. We not explored the role of imports into production from extra- regional suppliers, although we know these sources have grown relative to African suppliers. This aspect of research deserves further attention. Generally, the exploratory analysis presented in this Working Paper could be improved in a number of ways. Although we have described the theory which suggests differences in firm-level upgrading trajectories depending on the end market, and provided some empirical evidence which substantiates this view, case- study analysis would help to substantiate further the findings presented in this paper. Panel data analysis for countries with the available data may be more revealing. The distinction between intra- and extra- regional exporters for other regions included in the Global Enterprise surveys could also be undertaken in the future. Finally, the robustness of results could be further improved, and some variables excluded from analysis so as to sharpen the results. ODI Report 43 References AfDB/OECD/UNDP (2014), African Economic Outlook 2014: Global Value Chains and Africa’s Industrialisation. Paris: OECD Publishing for Organisation for Economic Co-operation and Development, African Development Bank, United Nations Development Programme. AfDB et al. (2014) Antras, P. and Helpman, E. (2004) Global Sourcing, Journal of Political Economy, Vol. 112, No.3. Available at: http://www.economics.harvard.edu/faculty/antras/papers/GSJPE.pdf Baldwin, R. and Venables, A. (2013) ‘Spiders and snakes: Offshoring and agglomeration in the global economy’. Journal of International Economics, Elsevier, vol. 90(2), pages 245-254. Bernard, A. B. and J. B. Jensen (2004), ‘Why some firms export’. Review of Economics and Statistics, 86: 561–569. Bernard, A. B., J. Eaton, J. B. Jensen and S. Kortum. 2003. .Plants and Productivity in International Trade. American Economic Review 93(4):1268.1290. Brandt, L., & Thun, E. (2010). The fight for the middle: upgrading, competition, and industrial development in China. World Development, 38(11), 1555-1574. Ciravegna, L. (2003). Global and regional integration of production in the Mercosur automotive value chains: the case of Fiat. DESTIN, London School of Economics. Clerides, S. Lach, S. and Tybout, J. (1998) "Is Learning By Exporting Important? Micro- Dynamic Evidence From Colombia, Mexico, And Morocco," The Quarterly Journal of Economics, MIT Press, vol. 113(3), pages 903-947, Farole and Winkler (2014) Making Foreign Direct Investment Work for Sub-Saharan Africa, World Bank: Washington DC. Available at: https://openknowledge.worldbank.org/bitstream/handle/10986/16390/978146480126 6.pdf?sequence=1 Fu, Xiaolan; Kaplinsky, Raphie and Kale, Dina (2011). Global manufactures prices: how do China’s exports compare? In: Fu, Xiaolan ed. China’s Role in Global Economic Recovery. Routledge Studies on the Chinese Economy. Abingdon, UK: Routledge, pp. 127–145. Gereffi, G., Humphrey, J. and Sturgeon, T. (2005) ‘The Governance of Global Value Chains’, Review of International Political Economy 12: 78-104. Goto, Kenta, Natsuda, Kaoru and Thoburn, John (2009) “Meeting the Challenge of China: The Vietnamese Garment Industry in the Post MFA Era”, mimeo Greenaway, D. and R. Kneller (2007), “Firm Heterogeneity, Exporting and Foreign Direct investment”, Economic Journal, Vol. 117, No. 517, pp. F134-F161(1), https://www.res.org.uk/economic/freearticles/february07.pdf ODI Report 44 Hoekman, B. and S. Jackson. (2013) “Reinvigorating the trade policy agenda: Think supply chain!” VoxEU, available at: http://www.voxeu.org/article/reinvigorating-trade- policy-agenda-thinksupply-chain. Humphrey, J.and O Memodovic. (2003) “The Global Automotive Industry Value Chain: What Prospects for Upgrading by Developing Countries?” Sectoral Studies Series. Vienna: United Nations Industrial Development Organization. Kamau (2009) Upgrading and technical efficiency in Kenyan garment firms: Does insertion in global value chains matter? PhD Thesis, IDS Nairobi. Kaplinsky, Raphael (2010). Asian drivers, commodities prices and the terms of trade. In: Nissanke, Machiko and Mavrotas, George eds. Commodities, governance and economic development under globalization. Basingstoke: Palgrave Macmillan. Kaplinsky, R. and A. Santos-Paulino (2005a), “A disaggregated analysis of EU imports: The implications for the study of patterns of trade and technology”, Cambridge Journal of Economics. Kaplinsky, R. and Wamae, W. (2010) The determinants of upgrading and value added in the African clothing sector : the contrasting experiences of Kenya and Madagascar, available at: http://www.open.ac.uk/ikd/documents/working-papers/ikd-working- paper-59.pdf Kaplinsky, R. and M. Morris (2001) A Handbook for Value Chain Research, Prepared for the International Development Research Centre (IDRC). Keane, J. (2014) The New GVC Discourse: What’s new, innovative, and missing? Briefing Paper, Overseas Development Institute: London. Keane, J. (2012) ‘The Governance of Glo bal Value Chains and the Transmission of the Global Financial Crisis to Producers in Africa and Asia’. Journal of Development Studies (48:6), pp 783-797. Keane, J., Cali, M. and Kennan, J. (2010) Impediments to Intra-Regional Trade in sub- Saharan Africa, paper prepared for the Commonwealth Secretariat. Available at: http://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion- files/7482.pdf Lee, J. and Chen, J. (2000) Dynamic Synergy Creation with Multiple Business Activities: Toward a Competence-based Growth Model for Contract Manufacturers in Sanchez, R. and Heene, A. (eds.) Research in Competence based Research, Advances in Applied Business Strategy Series Vol. C, JAI Press Markusen, J. and Venables, T. (2000) The Theory of Endowment, Intra-Industry and Multinational Trade, Journal of International Economics, Vol. 52, pp209-234. Melitz, M.J. 2003. "The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity," Econometrica, Econometric Society, vol. 71(6), pages 1695- 1725, November. Morris, M., Staritz, C., Plank, L. (2014) "Regionalism, end markets and ownership matter: Shifting dynamics in the apparel export industry in Sub Saharan Africa," Working Papers 46, Österreichische Forschungsstiftung für Internationale Entwicklung (ÖFSE) / Austrian Foundation for Development Research. Navas-Alemán, L. (2011) The Impact of Operating in Multiple Value Chains for Upgrading: The case of the Brazilian Furniture and Footwear Industries, World Development Vol. 39, No. 8, pp. 1386–1397, 2011 ODI Report 45 OECD and WTO (2013) ‘OECD -WTO Database on Trade in Value-Added First Estimates: 16 January 2013’. Paris: Organisation for Economic Co -operation and Development and World Trade Organization. Pietrobelli C and Saliola F 2008 Power relationships along the value chain: multinational firms, global buyers and performance of local suppliers, Cambridge Journal of Economics, 32(6), pp. 947-962. Porter, M. (1985) Porter, M., “Competitive Advantage: Creating and Sustaining Superior Performance”, Free Press: New York. Raikes, P., Jensen, M.F., and Ponte, S. (2000) Global commodity chain analysis and the French filiere approach: comparison and critique, Economy and Society 29(3), 390- 417. Shepard, B. and Stone, S. (2013) ‘Global Production Networks and Employment: A Developing Country Perspective’. OECD Trade Policy Papers, No. 154. Paris: the Organisation for Economic Co-operation and Development (http://dx.doi.org/10.1787/5k46j0rjq9s8-en). Staritz, C. and Morris, M. (2013) ‘Local Embeddedness, Upgrading and Skill Development: Global Value Chains and Foreign Direct Investment in Lesotho’s Apparel Sector’. Capturing the Gains Working Paper 20. Manchester: Capturing the Gains, University of Manchester (http://www.capturingthegains.org/pdf/ctg-wp-2013-20.pdf). Stevens, C., Keane, J. and Kennan, J. (2013) Supporting African Value Chains, paper prepared for the African Development Bank. Tijaja, J. (2012) Proliferation of Global Value Chains: Trade policy considerations for Indonesia. TKN Report December (available at: http://www.iheal.univ- paris3.fr/sites/www.iheal.univ-paris3.fr/files/global_value_chains_indonesia.pdf). Viner, J. (1950) The Customs Union Issue, Oxford University Press: Oxford. Wallerstein, M. (1974) The Modern World System: Capitalist Agriculture and the Origins of the European World Economy in the Sixteenth Century (New York: Academic Press). UNCTAD (2013) World Investment Report: Global Value Chains: Investment and Trade for Development. Geneva: United Nations Conference on Trade and Development (http://unctad.org/en/publicationslibrary/wir2013_en.pdf). WTO (2013) Global Value Chains in a Changing World, World Trade Organisation: Geneva. (http://www.wto.org/english/res_e/booksp_e/aid4tradeglobalvalue13_e.pdf). ODI Report 46 Appendix 1 Table A1: Review of Indicators of Interest Selected Indicators Category Indicators Economic Social Policy/Institutio GVC participation upgrading upgrading nal barriers related indicator Biggest Access to finance, Business Obstacle Access to land, licensing and Business licensing permits and permits, Customs and Corruption, Courts, trade regulations Crime, theft and Labour disorder, Customs regulations and trade regulations, Inadequately Electricity, trained workforce Inadequately trained workforce, Labour regulations, Political Instability, Practices of the informal sector, Tax administration, Tax rates, Transportation Finance Percent of firms Percent of firms Proportion of working identifying access to identifying access capital financed by finance as a major to finance as a supplier credit (%) constraint major constraint Proportion of working capital financed by supplier credit (%) Firm Proportion of private Proportion of private Characteri domestic ownership in foreign ownership in a stics a firm (%), Proportion firm (%) of private foreign ownership in a firm (%), Proportion of government/state ownership in a firm (%), Proportion of a firm held by the largest owner(s) (%), Percent of firms with legal status of publicly listed company, Percent of firms with legal status of privately held Limited Liability Company, Percent of firms with legal status of Sole Proprietorship, Percent of firms with legal status of Partnership, Percent of firms with legal status of Limited Partnership Innovation Percent of firms with Percent of firms Percent of firms with and an internationally- having their own an internationally- Technolog recognized quality Web site recognized quality y certification, Percent Percent of firms certification of firms using using e-mail to Percent of firms using technology licensed interact with technology licensed from foreign clients/suppliers from foreign companies*, Percent companies of firms having their own Web site, ODI Report 47 Percent of firms using e-mail to interact with clients/suppliers, Percent of firms with an annual financial statement reviewed by external auditors. Performan Capacity utilization Real annual Annual ce (%)*, Real annual sales growth employmen sales growth (%), (%) t growth Annual employment (%), growth (%), Annual Annual labour productivity labour growth (%) productivity growth (%) Regulation Senior management Days to obtain an s and time spent dealing import license Taxes with the requirements of government regulation (%), Number of visits or required meetings with tax officials, If there were visits, average number of visits or required meetings with tax officials, Days to obtain an operating license, Days to obtain a construction- related permit, Days to obtain an import license, Percent of firms identifying tax rates as a major constraint, Percent of firms identifying tax administration as a major constraint, Percent of firms identifying business licensing and permits as a major constraint Trade Days to clear direct Percent of firms Proportion of total exports through identifying inputs that are of customs, Percent of customs and foreign origin (%)* firms exporting trade regulations Percent of firms using directly or indirectly as a major material inputs and/or (at least 1% of sales), constraint supplies of foreign Percent of firms Days to clear origin* exporting directly (at imports from Proportion of total least 1% of sales), customs sales that are Proportion of total Days to clear exported directly (%) sales that are direct exports Proportion of total domestic sales (%), through customs sales that are Proportion of total exported indirectly sales that are (%) exported directly (%), Percent of firms Proportion of total exporting directly or sales that are indirectly (at least 1% exported indirectly of sales) (%), Days to clear Percent of firms imports from exporting directly (at customs*, Percent of least 1% of sales) firms using material inputs and/or supplies of foreign origin*, Proportion of total inputs that are of domestic origin (%)*, Proportion of total inputs that are of foreign origin (%)*, Days of inventory of main input*, Percent of firms identifying customs and trade ODI Report 48 regulations as a major constraint. Workforce Percent of firms Percent of Percent of firms offering formal firms identifying labour training, Proportion of identifying regulations as a workers offered an major constraint formal training (%)*, inadequatel Years of the top y educated manager's experience workforce working in the firm's as a major sector, Number of constraint permanent full-time Percent of workers, Number of firms temporary workers, offering Number of permanent formal production workers*, training Number of permanent Proportion non-production of workers workers*, Number of offered permanent skilled formal production workers*, training Number of permanent (%)* unskilled production Years of workers*, Proportion the top of unskilled workers manager's (out of all production experience workers) (%)*, working in Percent of firms the firm's identifying labour sector regulations as a major Number of constraint, Percent of permanent firms identifying an full-time inadequately workers educated workforce Number of as a major constraint temporary workers Proportion of unskilled workers (out of all production workers) (%)* ODI Report 49 Appendix 2 Figure A1: South Africa: top five imports (average 2011-13) 18,000 16,000 14,000 12,000 US$ million 10,000 8,000 6,000 4,000 2,000 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2709 oil (crude) 2710 oil (excl. crude) 8703 motor cars and vehicles 8525 transmission apparatus for radio-telephony 8471 automatic data-processing machines Source: Authors’ calculations using data from UN COMTRADE database. Share of total M = 29.8% ODI Report 50 Table A2: South Africa: suppliers of top five imports (average 2011-13) HS4 Description Supplier Share of total imports of product 2709 petroleum oils and oils obtained from Saudi Arabia 40.8% bituminous minerals, crude Nigeria 23.3% Angola 14.1% Iran 11.0% UAE 2.7% 2710 petroleum oils and oils obtained from India 25.5% bituminous minerals (excl. crude); preparations EU28 21.0% containing >= 70% by weight of petroleum oils Singapore 15.3% or of oils obtained from bituminous minerals, Mozambique 8.4% these oils being the basic constituents of the UAE 5.7% preparations, n.e.s.; waste oils containing mainly petroleum or bituminous minerals 8471 automatic data-processing machines and units China 59.3% thereof; magnetic or optical readers, machines EU28 14.6% for transcribing data onto data media in coded United States 4.3% form and machines for processing such data, Vietnam 3.7% n.e.s. Singapore 3.0% 8525 transmission apparatus for radio-telephony, China 33.8% radio-telegraphy, radio-broadcasting or EU28 18.0% television, whether or not incorporating Vietnam 12.4% reception apparatus or sound recording or India 9.5% reproducing apparatus; television cameras; still Mexico 9.3% image video cameras and other video camera recorders; digital cameras 8703 motor cars and other motor vehicles principally EU28 47.3% designed for the transport of persons, incl. Korea, Rep. 14.3% station wagons and racing cars (excl. motor India 11.9% vehicles of heading 8702) Japan 10.1% United States 9.1% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 51 Figure A2: South Africa: top five exports (average 2011-13) 12,000 10,000 8,000 US$ million 6,000 4,000 2,000 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 7110 Platinum 7108 Gold 2601 Iron ores and concentrates 2701 Coal 8703 Motor cars and vehicles Source: Authors’ calculations using data from UN COMTRADE database. Share of total X = 36.6% Table A3: South Africa: markets for top five exports (average 2011-13) HS4 Description Market Share of total exports of product 2601 iron ores and concentrates, incl. roasted iron China 68.6% pyrites EU28 11.8% Japan 10.0% Korea, Rep. 4.6% Singapore 2.0% 2701 coal; briquettes, ovoids and similar solid fuels India 26.3% manufactured from coal EU28 18.8% China 17.0% Taiwan 6.4% Israel 5.1% 7108 gold, incl. gold plated with platinum, unwrought Unspecified 99.8% or not further worked than semi-manufactured EU28 0.2% or in powder form 7110 platinum, incl. palladium, rhodium, iridium, Japan 33.8% osmium and ruthenium, unwrought or in semi- EU28 20.7% manufactured forms, or in powder form Switzerland 18.4% United States 16.8% Hong Kong 6.5% 8703 motor cars and other motor vehicles principally United States 45.4% designed for the transport of persons, incl. EU28 18.4% station wagons and racing cars (excl. motor Japan 10.0% vehicles of heading 8702) Namibia 5.9% Australia 4.8% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 52 Figure A3: Botswana: top five imports (average 2010-12) 2,500 2,000 US$ million 1,500 1,000 500 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 7102 Diamonds 2710 Oil (excl. crude) 8704 Motor vehicles for the transport of goods 8703 Motor cars and vehicles 2716 Electrical energy Source: Authors’ calculations using data from UN COMTRADE database. Share of total M = 38.7% Table A4: Botswana: suppliers of top five imports (average 2010- 12) HS4 Description Supplier Share of total imports of product 2710 petroleum oils and oils obtained from bituminous South Africa 94.9% minerals (excl. crude); preparations containing >= Namibia 2.8% 70% by weight of petroleum oils or of oils Mozambique 1.3% obtained from bituminous minerals, these oils Zimbabwe 0.7% being the basic constituents of the preparations, Switzerland 0.2% n.e.s.; waste oils containing mainly petroleum or bituminous minerals 2716 electrical energy South Africa 82.4% Namibia 8.4% Mozambique 5.4% Congo, Rep. 2.4% Zambia 0.8% 7102 diamonds, whether or not worked, but not EU28 70.6% mounted or set (excl. unmounted stones for pick- Namibia 10.5% up styluses, worked stones, suitable for use as South Africa 8.3% parts of meters, measuring instruments or other Israel 7.6% articles of chapter 90) India 1.1% 8703 motor cars and other motor vehicles principally South Africa 70.3% designed for the transport of persons, incl. Japan 21.1% station wagons and racing cars (excl. motor EU28 3.5% vehicles of heading 8702) Korea, Rep. 1.8% Singapore 1.7% 8704 motor vehicles for the transport of goods, incl. South Africa 68.1% chassis with engine and cab United States 16.2% EU28 6.2% India 4.8% Japan 2.5% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 53 Figure A4: Botswana: top five exports (average 2010-12) 900 5,000 800 4,500 700 4,000 3,500 600 US$ million 3,000 500 2,500 400 2,000 300 1,500 200 1,000 100 500 0 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 7501 Nickel mattes 7108 Gold 2603 Copper ores and concentrates 2836 Carbonates 7102 Diamonds (right-hand axis) Source: Authors’ calculations using data from UN COMTRADE database. Share of total X = 85.7% Table A5: Botswana: markets for top five exports (average 2010- 12) HS4 Description Market Share of total exports of product 2603 copper ores and concentrates South Africa 92.3% EU 6.5% Switzerland 1.2% Norway 0.0% China 0.0% 2836 carbonates; peroxocarbonates 'percarbonates'; South Africa 77.7% commercial ammonium carbonate containing Zimbabwe 15.6% ammonium carbamate Congo, Dem. Rep. 4.4% Zambia 2.2% Namibia 0.1% 7102 diamonds, whether or not worked, but not EU 84.1% mounted or set (excl. unmounted stones for pick- Israel 7.0% up styluses, worked stones, suitable for use as Switzerland 2.1% parts of meters, measuring instruments or other India 1.5% articles of chapter 90) United States 1.1% 7108 gold, incl. gold plated with platinum, unwrought South Africa 97.4% or not further worked than semi-manufactured Zambia 1.8% or in powder form United States 0.8% 7501 nickel mattes, nickel oxide sinters and other Norway 84.2% intermediate products of nickel metallurgy : Zimbabwe 14.7% South Africa 1.1% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 54 Figure A5: Namibia: top five imports (average 2011-13) 900 800 700 600 US$ million 500 400 300 200 100 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2710 Oil (excl. crude) 8703 Motor cars and vehicles 7102 Diamonds 2603 Copper ores and concentrates 8704 Motor vehicles for the transport of goods Source: Authors’ calculations using data from UN COMTRADE database. Share of total M = 27.0% Table A6: Namibia: suppliers of top five imports (average 2011- 13) HS4 Description Supplier Share of total imports of product 2603 copper ores and concentrates Switzerland 88.8% Peru 5.9% EU 5.2% 2710 petroleum oils and oils obtained from bituminous South Africa 74.8% minerals (excl. crude); preparations containing >= EU 8.6% 70% by weight of petroleum oils or of oils Tanzania 4.4% obtained from bituminous minerals, these oils Switzerland 3.9% being the basic constituents of the preparations, British Virgin Islands 2.4% n.e.s.; waste oils containing mainly petroleum or bituminous minerals 7102 diamonds, whether or not worked, but not Unspecified 44.1% mounted or set (excl. unmounted stones for pick- EU 31.0% up styluses, worked stones, suitable for use as Botswana 23.1% parts of meters, measuring instruments or other United States 1.2% articles of chapter 90) Israel 0.2% 8703 motor cars and other motor vehicles principally South Africa 89.9% designed for the transport of persons, incl. Botswana 3.6% station wagons and racing cars (excl. motor UAE 2.8% vehicles of heading 8702) EU 1.6% Japan 0.7% 8704 motor vehicles for the transport of goods, incl. South Africa 78.9% chassis with engine and cab United States 8.6% EU 6.3% UAE 2.5% China 2.0% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 55 Figure A6: Namibia: top five exports (average 2011-13) 1,600 1,400 1,200 1,000 US$ million 800 600 400 200 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 7102 Diamonds 2612 Uranium or thorium ores and concentrates 0303 Fish, frozen, excluding fish fillet 7901 Unwrought zinc 0304 Fish fillets and other fish meat Source: Authors’ calculations using data from UN COMTRADE database. Share of total X = 52.1% Table A7: Namibia: markets for top five exports (average 2011- 13) HS4 Description Market Share of total exports of product 0303 EU frozen fish (excl. fish fillets and other fish meat of 37.6% heading 0304) Congo, Dem. Rep. 20.4% South Africa 19.3% Mozambique 8.4% Angola 6.0% 0304 fish fillets and other fish meat, whether or not EU 88.9% minced, fresh, chilled or frozen South Africa 6.2% Australia 2.2% United States 1.9% Libya 0.1% 2612 uranium or thorium ores and concentrates Canada 36.9% EU 27.8% United States 20.0% China 14.8% Malawi 0.2% 7102 diamonds, whether or not worked, but not EU 43.1% mounted or set (excl. unmounted stones for pick- Botswana 27.7% up styluses, worked stones, suitable for use as Unspecified 17.1% parts of meters, measuring instruments or other United States 7.2% articles of chapter 90) Israel 2.5% 7901 unwrought zinc : EU 58.0% United States 15.1% Singapore 13.9% Switzerland 7.3% South Africa 4.3% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 56 Figure A7: Lesotho: top five imports (average 2008–9) 90 80 70 60 US$ million 50 40 30 20 10 0 2000 2001 2002 2003 2004 2008 2009 2710 Oil (excl. crude) 8702 Motor vehicles for the transport of >= 10 persons 9403 Other furniture and parts thereof 2711 Petroleum gases and other gaseous hydrocarbons 3304 Beauty, make-up, skin-care preparations Note: No data available for 2005–7. Graph excludes HS 9999 (UN Special Code) which, with a share of 6.7% of total import value, was the largest import (average 2008–9), as there are no data on composition or suppliers. Share of total M = 17.3% Source: Authors’ calculations using data from UN COMTRADE database. Table A8: Lesotho: suppliers of top five imports (average 2008– 9) HS4 Description Supplier Share of total imports of product 2710 petroleum oils and oils obtained from South Africa 100.0% bituminous minerals (excl. crude); preparations containing >= 70% by weight of petroleum oils or of oils obtained from bituminous minerals, these oils being the basic constituents of the preparations, n.e.s.; waste oils containing mainly petroleum or bituminous minerals 2711 petroleum gas and other gaseous hydrocarbons South Africa 100.0% 3304 beauty or make-up preparations and South Africa 100.0% preparations for the care of the skin, incl. sunscreen or suntan preparations (excl. medicaments); manicure or pedicure preparations 8702 motor vehicles for the transport of >= 10 Japan 41.0% persons, incl. driver EU 29.6% South Africa 20.1% United States 3.2% Gambia 2.3% 9403 furniture and parts thereof, n.e.s. (excl. seats South Africa 100.0% and medical, surgical, dental or veterinary furniture) Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 57 Figure A8: Lesotho: top five exports (average 2008-9) 400 350 300 250 US$ million 200 150 100 50 0 2000 2001 2002 2003 2004 2008 2009 6203 Men's suits, jackets, trousers etc., woven 8528 Television receivers 6110 Jerseys etc., knitted/crocheted 8538 Parts for electrical switching apparatus etc. 6204 Women's suits, dresses, skirts etc., woven Note: No data available for 2005–7. Share of total X = 46.8% Source: Authors’ calculations using data from UN COMTRADE database. Table A9: Lesotho: markets for top five exports (average 2008-9) HS4 Description Market Share of total exports of product 6110 jerseys, pullovers, cardigans, waistcoats and United States 63.5% similar articles, knitted or crocheted (excl. Canada 26.9% wadded waistcoats) South Africa 6.9% EU 1.1% Japan 0.9% 6203 men's or boys' suits, ensembles, jackets, United States 54.4% blazers, trousers, bib and brace overalls, South Africa 24.3% breeches and shorts (excl. knitted or crocheted, Canada 20.0% wind-jackets and similar articles, separate China 0.3% waistcoats, tracksuits, ski suits and swimwear) EU 0.3% 6204 women's or girls' suits, ensembles, jackets, United States 55.7% blazers, dresses, skirts, divided skirts, trousers, Canada 23.9% bib and brace overalls, breeches and shorts South Africa 19.8% (excl. knitted or crocheted, wind-jackets and EU 0.3% similar articles, slips, petticoats and panties, Japan 0.2% tracksuits, ski suits and swimwear) 8528 television receivers, whether or not incorporating South Africa 100.0% radio-broadcast receivers or sound or video recording or reproducing apparatus; video monitors and video projectors 8538 parts suitable for use solely or principally with South Africa 100.0% the apparatus of heading 8535, 8536 or 8537, n.e.s. Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 58 Figure A9: Swaziland: top five imports (average 2005–7) 300 250 200 US$ million 150 100 50 0 2000 2001 2002 2003 2004 2005 2006 2007 2710 Oil (excl. crude) 3302 Mixtures of odoriferous substances 8703 Motor cars and vehicles 1005 Maize (corn) 8704 Motor vehicles for the transport of goods Note: No more recent data available. Share of total M = 21.7% Source: Authors’ calculations using data from UN COMTRADE database. Table A10: Swaziland: suppliers of top five imports (average 2005–7) HS4 Description Supplier Share of total imports of product 1005 maize or corn South Africa 81.8% Zambia 18.2% 2710 petroleum oils and oils obtained from bituminous South Africa 98.8% minerals (excl. crude); preparations containing >= Taiwan 1.2% 70% by weight of petroleum oils or of oils obtained from bituminous minerals, these oils being the basic constituents of the preparations, n.e.s.; waste oils containing mainly petroleum or bituminous minerals 3302 mixtures of odoriferous substances and mixtures, South Africa 100.0% incl. alcoholic solutions, based on one or more of these substances, of a kind used as raw materials in industry; other preparations based on odoriferous substances, of a kind used for the manufacture of beverages 8703 motor cars and other motor vehicles principally South Africa 29.8% designed for the transport of persons, incl. station Mozambique 25.0% wagons and racing cars (excl. motor vehicles of Japan 24.4% heading 8702) EU28 20.7% 8704 motor vehicles for the transport of goods, incl. South Africa 89.0% chassis with engine and cab Japan 8.4% EU28 2.4% Ghana 0.2% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 59 Figure A10: Swaziland: top five exports (average 2005-7) 700 600 500 US$ million 400 300 200 100 0 2000 2001 2002 2003 2004 2005 2006 2007 3302 Mixtures of odoriferous substances 1701 Cane or beet sugar 3824 Prepared binders for foundry moulds 4703 Chemical wood pulp, soda or sulphate 6109 T-shirts etc., knitted or crocheted Note: No data available for 2005–7. Share of total X = 67.1% Source: Authors’ calculations using data from UN COMTRADE database. Table A11: Swaziland: markets for top five exports (average 2005–7) HS4 Description Market Share of total exports of product 1701 cane or beet sugar and chemically pure sucrose, EU28 49.3% in solid form Uganda 43.1% South Africa 5.0% New Zealand 2.6% 3302 mixtures of odoriferous substances and mixtures, Australia 39.1% incl. alcoholic solutions, based on one or more of Zimbabwe 32.0% these substances, of a kind used as raw materials South Africa 28.9% in industry; other preparations based on odoriferous substances, of a kind used for the manufacture of beverages 3824 prepared binders for foundry moulds or cores; South Africa 63.2% chemical products and preparations for the Mozambique 36.8% chemical or allied industries, incl. mixtures of natural products, n.e.s. 4703 chemical wood pulp, soda or sulphate (excl. South Africa 100.0% dissolving grades) 6109 t-shirts, singlets and other vests, knitted or United States 44.1% crocheted Mozambique 22.5% Lesotho 21.1% South Africa 12.3% Source: Authors’ calculations using data from UN COMTRADE database. ODI Report 60 Appendix 3: Summary of Enterprise Survey Analysis Table A12: Ownership and Direct Exporting SACU Firms Country All foreign Non- Exporters Foreign and ownership exporters exporting South Africa 137 933 124 124 Namibia 28 88 19 11 Botswana 129 237 31 18 Lesotho 50 123 28 28 Swaziland 27 45 25 16 Total 371 1425 227 197 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A13: Firm Age SACU Sample Non- Country All Firms Exporters Exporters Non- Exporting Age (Av) Age (Av) Exporter Firms South Africa 933 124 27 23 Namibia 88 19 26 20 Botswana 237 31 23 20 Lesotho 123 28 15 18 Swaziland 45 25 19 20 Total/Av. 1425 227 24.3 22.1 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A14: Firm Size SACU Sample Country Exporters Non-Exporters Av. Size Av. Size South Africa 146 92 Namibia 199 23 Botswana 115.8 73 Lesotho 2646 107 Swaziland 412.2 57 Total/Av. 474.5 84.6 Source: World Bank Enterprise Surveys for each country, described in Section Two. ODI Report 61 Table A15: Licences and International Quality Certification Foreign International Quality Certification Licence Country Exporter Dummy % % Don't In No Yes Licence No Yes Cert. know Process Non- 0 0 0 - 5 177 4 50 21.2 Botswana Exporter Exporters 1 0 0 - 0 21 1 9 29.0 Non- 0 0 0 - 3 97 0 23 18.7 Lesotho Exporter Exporters 1 0 0 - 1 23 0 4 14.3 Non- 0 71 16 18.4 1 70 0 17 19.3 Namibia Exporter Exporters 1 15 4 21.1 0 6 0 13 68.4 Non- Republic of 0 470 76 13.9 0 554 0 248 30.9 Exporter South Africa Exporters 1 100 24 19.4 0 79 0 45 36.3 Non- 0 40 5 11.1 0 39 0 6 13.3 Swaziland Exporter Exporters 1 20 5 20.0 0 11 0 14 56.0 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A16: Length of relationship with supplier, and sub- contracting Sub-contracting Arrangements Av. no. of years % Firms Country Firm Type firm has known Sub- No- Don’t Sub- supplier Total Contract Subcontract Know Contracte d South Non-Exporter 13 77 469 0 546 14 Africa Exporter 11 22 102 0 124 18 Namibia Non-Exporter 9 8 79 0 87 9 Exporter 9 2 17 0 19 11 Botswana Non-Exporter 14 60 175 2 237 25 Exporter 13 13 18 0 31 42 Lesotho Non-Exporter 9 26 82 5 113 23 Exporter 8 6 21 1 28 21 Swaziland Non-Exporter - 9 39 0 48 19 Exporter - 12 13 0 25 48 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A17: Manager’s experience in SACU firms Managers Experience Mean Count 0 18 237 Botswana Export Dummy 1 19 31 0 13 123 Lesotho Export Dummy 1 12 28 0 11 87 Namibia Export Dummy 1 16 19 Republic of 0 13 933 Export Dummy South Africa 1 16 124 0 8 45 Swaziland Export Dummy 1 9 25 Source: World Bank Enterprise Surveys for each country, described in Section Two. ODI Report 62 Table A18: Managers’ education level in exporting firms None / % Tertiary Tertiary Vocational Don't Total Educ. and and above / Other know above Export 0 163 71 3 237 69 Botswana Dummy 1 27 4 0 31 87 Export 0 65 52 6 123 53 Lesotho Dummy 1 11 16 1 28 39 Export 0 44 43 0 87 51 Namibia Dummy 1 9 10 0 19 47 Republic of South Export 0 448 474 11 933 48 Africa Dummy 1 74 48 2 124 60 Export 0 23 21 1 45 51 Swaziland Dummy 1 19 6 0 25 76 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A19: Workforce skills and development Proportion offering Don't Total No. No Yes Formal Training Know Firms (%) 0 0 109 128 237 54 Botswana Export Dummy 1 1 13 17 31 55 0 3 55 65 123 53 Lesotho Export Dummy 1 0 15 13 28 46 0 0 54 33 87 38 Namibia Export Dummy 1 0 5 14 19 74 Republic of 0 11 566 356 933 38 Export Dummy South Africa 1 0 55 69 124 56 0 0 29 16 45 36 Swaziland Export Dummy 1 0 6 19 25 76 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A20: Workforce Education: An obstacle for SACU firms? % Firms Major Minor Moderate No Severe Total (Severe/Major) Export 0 57 44 52 62 22 33.3 237 Botswana Dummy 1 9 5 9 4 4 41.9 31 Export 0 10 26 21 57 9 15.4 123 Lesotho Dummy 1 5 5 6 10 2 25.0 28 Export 0 21 11 23 25 7 32.2 87 Namibia Dummy 1 4 1 7 5 2 31.6 19 Republic of Export 0 57 164 114 583 15 7.7 933 South Africa Dummy 1 10 35 22 52 5 12.1 124 Export 0 7 8 7 19 4 24.4 45 Swaziland Dummy 1 5 2 10 6 2 28.0 25 Source: World Bank Enterprise Surveys for each country, described in Section Two. Table A21: Labour Market Regulations: An obstacle? % Firms No Minor Moderate Major Severe Total (Severe/Major) Export 0 85 59 54 31 8 16 237 Botswana Dummy 1 10 10 9 1 1 6 31 Export 0 61 29 20 8 5 11 123 Lesotho Dummy 1 16 5 4 2 1 11 28 Export 0 44 16 15 10 2 14 87 Namibia Dummy 1 5 5 3 4 2 32 19 Republic of Export 0 629 167 88 41 8 5 933 South Africa Dummy 1 67 31 14 11 1 10 124 Export 0 26 11 4 3 1 9 45 Swaziland Dummy 1 5 7 4 6 3 36 25 Source: World Bank Enterprise Surveys for each country, described in Section Two. ODI Report 63 Table A22: Customs and Trade Regulations: An Obstacle? No Minor Moderate Major Severe % Severe/Major Total Botswana Export 0 113 49 40 27 8 15 237 Dummy 1 5 4 10 9 3 39 31 Lesotho Export 0 56 24 22 13 8 17 123 Dummy 1 12 4 5 5 2 25 28 Namibia Export 0 47 16 18 6 0 7 87 Dummy 1 11 3 2 3 0 16 19 Republic of Export 0 787 94 27 19 6 3 933 South Africa Dummy 1 103 15 4 1 1 2 124 Swaziland Export 0 32 4 6 3 0 7 45 Dummy 1 10 6 4 2 3 20 25 Source: World Bank Enterprise Surveys for each country, described in Section Two. ODI Report 64 Appendix 4 Global Enterprise Survey Comparison Table A23: Global Enterprise Survey Data Regions All Foreign All Non- All Foreign and Ownership Exporters Exporter Exporting s AFR 3,302 18,785 1,456 549 EAP 980 7,398 1,246 429 ECA 23,382 21,755 3,673 773 LAC 22,721 21,735 3,852 971 SAR 213 5,782 970 74 Total 50,598 75,455 11,197 2,796 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. Table A24: Ownership and Exporting Regions Ownership Type Domestic Foreign Foreign Domestic Non- Exporter Non- Exporter exporter exporter AFR 3,302 16,939 16,032 907 2,753 549 EAP 980 7,664 6,847 817 551 429 ECA 2,046 23,382 20,482 2,900 1,273 773 LAC 2,866 22,721 2,881 19,840 1,895 971 SAR 213 6,539 5,643 896 139 74 Total 9,407 77,245 51,885 25,360 6,611 2,796 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. ODI Report 65 Table A25: Firm Age: Global Enterprise Survey Region All Firms Exporters Non- Exporters Non-exporters Exporters Age(AV) Age(Av) AFR 18,463 1,437 26.71 18.99 EAP 7,260 1,231 19.36 18.10 ECA 21,492 3,628 23.036 17.718 LAC 21,497 3,831 34.03 28.481 SAR 5,744 967 22.80 21.59 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. Table A26: Firm Size Global Enterprise Survey Results Region All Firms Exporters Non-Exporters Non-exporters Exporters Av.Size Av.Size AFR 18592 1443 261.68 41.33 EAP 7382 1244 501.44 104.765 ECA 21604 3649 175.86 77.08 LAC 21658 3847 289.29 94.41 SAR 5763 969 517.01 75.23 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. Table A27: Managers Experience Regions Average Yr. AFR Export 0 13.56 Dummy 1 13.08 EAP Export 0 14.65 Dummy 1 16.48 ECA Export 0 15.45 Dummy 1 17.96 LAC Export 0 20.85 Dummy 1 21.45 SAR Export 0 17.52 Dummy 1 19.51 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. ODI Report 66 Table A28: International Quality Certification Enterprise Survey Region Exporter International Quality % Certified Dummy Certification Don’t know No In progress Yes AFR Non- 0 235 15842 201 2507 13.34 Exporter Exporters 1 19 788 27 622 42.719 EAP Non- 0 101 5230 71 1996 26.98 Exporter Exporter 1 17 533 29 667 54.33 ECA Non- 0 298 17087 293 4077 18.74 Exporter Exporters 1 48 1785 95 1745 47.50 LAC Non- 0 236 17530 463 3506 16.13 Exporter Exporters 1 28 2017 135 1672 43.40 SAR Non- 0 32 5043 30 677 11.70 Exporter Exporters 1 10 585 19 356 36.70 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. Table A29: Workforce Education: An obstacle? Regional Results Regions Minor Moderate Major Sever No % Firms Total (Severe/Major) AFR Export 0 4415 2585 1251 503 10031 9.33 18785 Dummy 1 363 297 151 58 587 14.35 1456 EAP Export 0 1584 674 224 45 4871 3.63 7398 Dummy 1 320 150 66 17 693 6.66 1246 ECA Export 0 3603 3149 1266 509 13228 8.15 21755 Dummy 1 792 754 284 136 1707 11.43 3673 LAC Export 0 4307 5777 3012 1684 6955 21.605 21735 Dummy 1 764 1163 670 268 987 24.35 3852 SAR Export 0 1447 859 278 103 3095 6.58 5782 Dummy 1 308 209 49 17 387 6.80 970 Source: Global Enterprise Surveys. Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. Table A30: Regional Results for Labour Market Regulations Regions Minor Moderate Major Sever No % Firms Total (Sever/Major) AFR Export 0 3555 2697 2137 1129 9267 17.38 18785 Dummy 1 296 290 278 150 442 29.39 1456 EAP Export 0 1132 490 287 112 5377 5.3 7398 Dummy 1 340 197 82 19 608 8.10 1246 ECA Export 0 2136 2097 1229 787 15506 9.26 21755 Dummy 1 695 621 350 225 1782 15.65 3673 LAC Export 0 3118 3462 1844 1158 12153 13.81 21735 Dummy 1 877 1174 532 332 937 22.42 3852 SAR Export 0 1316 1080 506 192 2688 12.07 5782 Dummy 1 291 293 109 29 248 14.22 970 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. ODI Report 67 Table A31: Regional Results for Customs and Trade Regulations Regions Minor Moderate Major Severe No % Firms Total (Severe/Major) AFR Export 0 4,108 2,808 2,281 1,159 8,429 18.31 18,785 Dummy 1 339 292 202 102 521 20.87 1,456 EAP Export 0 1,873 1,079 542 151 3,753 9.36 7,398 Dummy 1 367 205 85 15 574 8.02 1,246 ECA Export 0 3,313 3,727 3,542 1,892 9,281 24.97 21,755 Dummy 1 634 761 661 386 1,231 28.50 3,673 LAC Export 0 3,814 5,792 5,025 2,385 4,719 34.09 21,735 Dummy 1 650 1,173 998 417 614 36.73 3,852 SAR Export 0 1,480 1,349 647 275 2,031 15.94 5,782 Dummy 1 256 311 167 37 199 21.03 970 Source: Global Enterprise Surveys Note: the total number of firms differs to those used in our analysis sue to missing data/coding issues as explained in Section 2. ODI Report 68 ODI is the UK’s leading independent think tank on international development and humanitarian issues. Our mission is to inspire and inform policy and practice which lead to the reduction of poverty, the alleviation of suffering and the achievement of sustainable livelihoods. We do this by locking together high-quality applied research, practical policy advice and policy- focused dissemination and debate. We work with partners in the public and private sectors, in both developing and developed countries. Readers are encouraged to reproduce material from ODI Working Papers for their own publications, as long as they are not being sold commercially. As copyright holder, ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the author(s) and do not necessarily represent the views of ODI. © Overseas Development Institute 2013. This work is licensed under a Creative Commons Attribution-NonCommercial Licence (CC BY-NC 3.0). ISSN (online): 1759-2917 ISSN (print): 1759-2909 Cover image: Overseas Development Institute 203 Blackfriars Road London SE1 8NJ Tel +44 (0)20 7922 0300 Fax +44 (0)20 7922 0399