ࡱ> q` I bjbjqPqP 6::K(( :::8r|"6D(lll0x """""""$d$h&\3"{3"llH"0nnnNll "n "nn9Xql @6<:2 (dx"H"('<('Pq('q8 nC3"3"j",Er@,r April 2008, Draft. Globalisation and Deep Integration Michael Gasiorek Peter Holmes (corresponding author: p.holmes@sussex.ac.uk) The Centre for the Analysis of Regional Integration at 鶹ӳ This paper was written following a series of workshops and seminars on deep integration held at 鶹ӳ University over the course of the last year. We would therefore like to acknowledge the extremely helpful comments received and discussions had with colleagues at 鶹ӳ including David Evans, Leo Iacovone, John Humphrey, Sherman Robinson, Jim Rollo, Alasdair Smith, Alan Winters, and Dirk Willenbockel. We would also like to thank Ahmed Ghoneim for his insightful comments on earlier drafts. Of course the usual disclaimers apply and responsibility for the views expressed in this paper is that of the authors. Contents:  TOC \o "1-3" \h \z \u   HYPERLINK \l "_Toc197161446" 1. Introduction  PAGEREF _Toc197161446 \h 2  HYPERLINK \l "_Toc197161447" 2. Deep institutional integration  PAGEREF _Toc197161447 \h 3  HYPERLINK \l "_Toc197161448" 3. Deep market integration  PAGEREF _Toc197161448 \h 6  HYPERLINK \l "_Toc197161469" 4. The Gains from Deep Integration  PAGEREF _Toc197161469 \h 13  HYPERLINK \l "_Toc197161470" 5. Achieving deeper institutional integration  PAGEREF _Toc197161470 \h 17  HYPERLINK \l "_Toc197161471" 6. Conclusion  PAGEREF _Toc197161471 \h 22  Introduction The world economy is currently undergoing substantive changes which are frequently referred to in the context of the term globalisation. These substantive changes are being driven in part by the decline in transport and communication technologies, in part by the greater integration of economies such as China, India, Brazil, and Korea into the world economy, and in part by the ongoing process of multilateral and regional integration. The changes appear to be transforming the very nature of inter and intra-firm relationships and which are then manifesting themselves in the changing patterns of trade, investment, and production. In this paper we argue that the notion of deep integration is central to understanding the transformations which are taking place. Deep integration is a term developed by R.Z. Lawrence (1996) which is typically used in a range of contexts and with its meaning not always clearly specified. It is frequently used in the context of regional trade agreements where the focus is on integration processes which deal with behind the border barriers. However, often when used in this context there quite what the behind the border issues are, or what it means to achieve deep integration are poorly specified. The aim of this paper is to understand the role of deeper integration in the context of the increasingly globalised world economy. This is important not simply in terms of understanding the changing patterns of production, trade and investment, but also in terms of considering optimal policy responses. In this paper we therefore, first consider carefully what is meant by the term deep integration. We argue that deep integration can be seen both as a process driven by both public and private policy actions, and in terms of outcomes with respect to the degree and nature of firm level interdependence. The former can be characterised as deep institutional integration, and the latter as deep market integration. Secondly, we consider the possible impact of deeper integration where we highlight what the nature of the potential gains may be gains from such integration. Thirdly, and on the basis of the preceding, we identify ways in which deep integration might be achieved in particular in the context of regional trade agreement. In the first instance deep integration can be thought of as being defined by a process be this public or market let. In this sense of the terms it refers to the removal of barriers to trade that are domestic or behind the border - normally domestic rules and regulations that may affect both home and foreign firms - but which also restrict foreign access. Barriers to trade in services, barriers to investment, the absence of an effective competition policy are all of this kind. Since some industries are allowed to practice self regulation via private rules and norms we should include private measures as well as public ones. Evaluating the extent of this aspect of deep integration would call for an assessment of the different types of measures undertaken by governments and firms. A second possible meaning of the term concerns the outcome of the process and in particular the degree and nature of firm level interaction within and across borders both with other firms and with their own subsidiaries. The rise of vertical fragmentation, or trade in tasks is a reflection of this process of deeper integration between firms and states. As opposed to looking at measures undertaken evaluating the extent of this aspect of deep integration can only be done by finding suitable measures of the changing nature of firm level relations, and of patterns of trade derived from these relations. This could be achieved by looking at a number of indicators - institutional (eg prevalence of subcontracting), home market preference, price convergence, measures of intra-industry trade, measures of vertical specialisation. Intra-industry trade (IIT) indicators appear to capture two aspects of deep integration in outcome terms, namely fine degrees of either horizontal or vertical specialisation. IIT measures identify both niche specialisation within product varieties, whether of higher or lower quality or just different, and trade flows where different components are made in different places. For the purposes of this paper, where we need to distinguish between the process and the outcome, than we use the term deep market integration (DMI) for the outcomes, and deep institutional integration (DII) for the process. It is our hypothesis - not an assumption - that deep institutional integration is likely to lead to deep market integration, but that deep market integration is also create a demand for further institutional integration. This interaction inevitably complicates empirical analysis. Deep institutional integration In principle a distinction can be made between measures applied at the border such as tariffs or quantitative restrictions which are enforced at the border by customs officials on the one hand and on the other domestic regulatory measures which affect goods (after they have passed the borders and are then circulated in the internal market) at the point of sale. Deep institutional integration typically refers to a process in which it is the domestic economic (regulatory) policies/measures, which are modified in order to remove barriers to trade. These can include issues such as customs procedures, regulation of domestic services production that discriminate against foreigners, product standards that differ from international norms or where testing and certification of foreign goods is complex and perhaps exclusionary, regulation of inward investments, competition policy, intellectual policy protection and the rules surrounding access to government procurement. This is by contrast with shallow integration under which it is the border measures essentially tariffs and quotas, which are eliminated. These concepts are also linked to the distinction between positive and negative integration. Negative integration means the prohibition of certain measures that will restrict trade. Positive integration is an obligation to adopt certain common policies. These distinctions become somewhat blurred in practice. The negative integration prohibitions may extend to domestic rules, e.g. on subsidies and a sufficiently broad set of restrictions on policy may end up prescribing what may be done! For example a trade agreement may state that signatories may not ban the sale of products produced to international standards. Cast in the form of a ban, a prohibition rather than a positive obligation, it nevertheless creates a positive obligation to allow the use of international standards. In order to make more systematic the analysis of deep institutional integration we need to first detail the types of both border and behind the border barriers to trade, which might exist; and then consider the ways in which these might be legitimately considered as barriers to deeper integration. A framework for this analysis is laid out in Table 1. In the table we look at the different ways in which different kinds of obstacles to trade may impinge on trade. The first column of the table details the types of barrier, which might exist, and then the subsequent columns provide specify the nature of both border and behind the border impediments which may occur with respect to that barrier. The last column is the area where we think deep integration measures as identified earlier may be able to assist. When considering the border and behind the border barriers, it is also useful to distinguish between barriers which are directly related to physical infrastructure, and barriers which arise from institutional/regulatory policy measures undertaken. Hence, if we take the first row of the table, here we identify tariffs and quotas. These are clearly in the first instance border barriers, which may limit trade directly because of the actual policy measure itself (the height of the tariff, or the amount of the quota), and as a result of customs clearance procedures at the point of entry. There could also be behind the border barriers arising from the procedures which may need to be undertaken in order to be able to supply the good. Here we seen that even for goods the distinction between deep and shallow integration is sometimes blurred, and this may be even more so once one goes beyond tariffs and quantitative restrictions that affect only imports. In looking at technical barriers to trade (TBTs) and Sanitary and Phytosanitary (SPS) measures affecting food safety it is important to distinguish between: standards which are in principle voluntary, mandatory technical regulations, conformity assessment (CA) by which compliance with standards or regulations is checked, and accreditation in which the inspection system for CA is itself assessed. Here we see that there are both border barriers which may arise because of the e.g. warehousing infrastructure, or the policy that are in place for the inspection of goods; and behind the border barriers which are concerned with the presence of testing and certification laboratories, and the measures that are in place for conformity assessment and accreditation, as well as the presence or not of both private and public standards which in turn may or may not require appropriate legislation. The absence of those standards, or the absence of a capacity to prove that given standards may have been met will clearly impede the capacity for firms to supply the market for which those standards are required. Hence, introducing standards, conformity assessment, and accreditation provides a clear example of deep integration which then serves to facilitate trade. For services obstacles to international transactions are generally internal regulations although some services do require services or their providers to cross frontiers. Shallow integration proposals can reduce the impact of the barriers that come from rules applied at the border, but deep integration requires one to address domestic processes. In both cases however modulating formal regulatory barriers will not facilitate trade unless the physical or human infrastructure is improved. For example liberalizing rules of public procurement will not help if the process for making bids still requires a heavy resource commitment. Anti-dumping duties are levied at borders but they have a domestic element in that there may be scope for appeals and rebates all of which are internal. Anti-competitive practices and their control are usually within the national economy but monopolised distribution channels might start to bite at the frontier itself. Table 1. Understanding Barriers to Deep Integration Border barrier aspectsBehind the border dimensionInfrastructureInstitutional / regulatory / policyInfrastructureInstitutional / regulatory / policyTariffs / quotasCustoms clearanceActual measures eg. tariffs, quotasDomestic implementation of applications for clearances etc especially for quotasStandards: (SPS, TBT)Warehouse facilitiesInspection processesLabsStandards, regulations etcInvestmentn/an/aAccess to facilitiesInvestment rulesIPR Checks to secure compliancePatent officeIPR legislation and rulesTrade defence (AD, CVDs Safeguards) Customs clearanceActual measuresn/aBureaucratic procedures for imposing duties and appealing etcServices: e.g. financial, insurance, transport, telecoms)InternetPossible blockages on internetCore network for peripheral servicesRegulationsGovt procurementXPurchasing systemLegislation and rulesComp. policy private firmsAny marketing obstacles by incumbentsCourtsLegislation and rulesComp policy- state aidsAids regime and ways to challenge On the basis of the preceding we can then identify a checklist of factors that are usually associated with deep integration, and which are increasingly frequently included in RTA discussions. These typically include: Investment rules. A degree of regulatory harmonisation (product standards or process standards) with an approach to Harmonisation, National Treatment, or Mutual Recognition: This is a key issue: the EU approach unlike NAFTA has been to stress the need for harmonisation at least of minimum standards. Here the clear trade off is between the domestic costs and the prospect of market access Provisions on anti-dumping/countervailing duties Provisions on subsidies Competition policy alignment? Here there is much dispute about. For example for LDCs there is a need is to ensure that cost of compliance is not disproportionate and that there is not an asymmetry of obligations Services schedule relative to GATS commitments, especially rules on movement of natural persons. Since, in many cases, GATS does not impose serious constraints, systematic evaluation of costs and benefits benefits of mutual market access may depend on provisions such as visa rules for temporary workers that may be excluded from the RTA. Harmonisation etc of issues going beyond even behind border concerns , e.g. general legal norms other than TBTs, etc. Institutional framework necessary to ensure that agreements are credible and enforceable. This may include some form of supra-national rule making system, an autonomous secretariat, and/or an ex post binding dispute settlement. Political integration / political benefits / non-trade political conditionality? In accompanying work in collaboration with A.Ghoneim we have sought to apply this checklist to a number of agreements including the EU-Egypt Association agreement. We will not explore this here but proceed to ask the circumstances under which one would wish to include these elements. The core reason is we assume that the removal of tariffs alone will leave other institutional barriers unaffected. Deep market integration Much of the row entries in Table 1 above directly refer to the process of public or state led deep institutional integration. From the point of view of firms, changes in policy in any of these areas potentially raises the possibility of deeper market integration. This possibility of deeper market integration arises in two principal ways. First, it increases the contestability of markets by making it easier for firms to access each others market and on a more even or level playing field. Hence action with regard to competition policy, government procurement, trade defence or investment rules are all likely to have this impact. Secondly, more indirectly, and in particular in conjunction with integration with respect to standards it allows for greater interconnectedness between firms. This greater interconnectedness may be between different firms, or between firms and their subsidiaries and is thus likely to occur by dividing up tasks along different parts of the value chain. This occurs both with regard to manufacturing and service activities, and in the linkage between the two. In considering deep integration it is then clear that standards play an important role. Significantly deeper integration with respect to standards is something which can be achieved either by public or private agents. It is also important to note that standards can be either barriers to trade or facilitators of trade, and it is difficult to identify beforehand their specific nature. They can be drawn up nationally, regionally, or multilaterally, or by the private sector and be mandatory or voluntary. In the case of private standards they may be made mandatory by the market and can be both a challenge and an opportunity for developing country exporters. An important issue with regard to standards is that where they matter (and their remit is constantly widening) international trade is increasingly based on quality and technology rather than simply on price. This claim is something of a clich but we seek here to give it a firm basis. For an increasing number of commodities public and private standards (in the broad sense of this term) are creating a situation in which the notion of a trade off between price and quality no longer applies. That is to say, price competition enters into the game only if certain minimum standards have been met. This means that in such markets there is no positive price at which certain low quality items can be sold. Such a concept runs against the orthodox precepts of neo-classical economics but is immediately intelligible when one allows for uncertainty and transactions costs. In a world of perfect information and no transactions costs, standards would not be necessary. Buyers would be able to costlessly assess the quality of goods and services and evaluate and predict the adverse cost of lower quality. In real life consumers do not know the consequences of certain health or breakdown risks and are willing to pay a premium for standardized commodities. Moreover there are certain risks of breakdown that one would not take at any price. For those who can comply, externalities and increasing returns to scale from common standards may actually make the marginal cost of standardisation negative. We can then identify what might be termed Smithian as opposed to Ricardian trade. In Ricardian trade commodities are homogenous, or at least the quality is instantly recognisable and measurable and differs only in a quantitative way, such as in the percentage content of a certain material. In this model profit margins will not be related to quality and comparative advantage is based on cost of inputs and cost-efficiency alone. In the Smithian model, the assumption is that there are economies of scale and the benefits of specialisation come from chopping up the production chain. But of course as the processes are separated the outputs of one process become the inputs to another and this requires some sort of coordination mechanism between suppliers and users-consumers. There was extensive literature on transactions costs (including Williamson, 1975) which argued that the increase in transaction costs was overcome by internalizing them through vertical integration or through reinforced and cross-ownership structures with a consequent expansion intra-firm trade. Williamson argued that even between firms in the same town, the problems of monitoring quality were likely to be so great that only hierarchical control of production processes could ensure quality. Hence, that the factory system arose not from technological economies of scale but from the need for monitoring of quality. For example if a clothing manufacturer could only see the outside of bales of cloth he would not know what they were like throughout: he would prefer to have control over the workplace where they were made. External producers were marginalized. However, new management and information systems have revitalised the subcontracting and outsourcing relationship. Best (1990) took a complementary approach arguing that the development of technical standards in the US in the 19th century both directly raised productivity across the whole economy (because US standard screws has very well shaped threads) but also generated huge externalities as the armys demand that parts for rifles be interchangeable wherever they had been made. The result was that firms and workshops could specialise in very fine lines of activity and gain economies of scale and learning effects. The United States thus achieved true deep market integration. A viable standards system is a way to reduce the transactions costs of unreliable and potentially incompatible components. However standards without a guarantee of quality are not adequate. The entire outsourcing movement is based on the need to find ways round the Williamson problem. The Japanese car industry has often been seen as a model for others to follow, in two respects, first the striving for zero defects and the extensive use of external supply chains. Enthusiasts for the Toyota model originally based on the quality assurance ideas of the American engineer Deming, showed that there is no trade-off between physical productivity and production quality: this is because high productivity is best achieved by ensuring that all output coming off the production line can be sold without further modification. Japanese firms have been very successful in both cars and especially in electronics in extending this model to suppliers across the whole of East Asia, despite the lack of institutional support. John Suttons work has vividly illustrated the role of quality standards in the general sense. In an investigation of how the Indian and Chinese car component industries have developed he shows that those firms who have become successful subcontractors have done so by reducing the rejection rates of faulty products to levels comparable to US, EU or Japanese suppliers. Those who have not done so have not been able to enter the value chains. The reason why it is necessary to use sophisticated quality assurance processes rather than simply relying on monitoring output quality is that the defects in the intermediate output may not be easily visible. One component proving defective could damage an entire production line. Similar considerations apply to consumer goods, especially food products which are increasingly being bought and sold in the same way as intermediate industrial products, in that supermarkets are imposing increasingly tight quality standards treating them as inputs into a production process. Necessary does not however mean sufficient: producers who cannot meet these quality standards will not be able to do so merely by the creation of a regional monitoring capacity, though this is clearly the first step in improving quality. The phenomena we are analysing here leads to deeper market integration, and have to do with intra-industry trade in two ways, which we define as horizontal Smithian trade and vertical Smithian trade. These are nothing more than the two well known types of intra-industry trade. The first type where producers market a finished product that fits into a highly differentiated niche where reputation, brand and quality allow a price premium to be obtained that cannot be eroded by new entry. This process is sustained by what the economists calls love for variety, a typical feature characterizing all the so-called new trade theory models as well as recent model of industrial organization and heterogeneous firms. In this type of trade, standards work as amplifiers and catalysts by allowing the creation of recognizable brands and types, developing new niches that consumers can identify without incurring in screening and search costs. Consider for example the expansion of organic market and related standards or the fair trade market. The second type of intra industry trade is where the value chain is broken up, and in Adam Smiths example different parts of the pin production process are located all over the world. For this to be able to happen we need mechanism for the contracts between upstream and downstream producers to be able to be very carefully and reliably monitored and enforced. This depends on the capacity of the producers to guarantee quality, which depends both on the market and on national and global public as well as private provision of standards facilities. In such markets low wages cannot offer an alternative to compliance (cf Jaffee and Henson). This type of trade is frequently variously referred to in the literature as vertically fragmented trade, or more recently trade in tasks. The term trade in tasks emphasises the point that this type of intra-industry trade may not simply be in manufacturing but is likely to cross the boundaries between manufacturing and services. In Table 2 we summarise the role of standards in the new types of specialization. The key point is that intra-industry trade of both kinds allows producers to specialise in a particular product or process for which a premium price can be extracted and specific expertise gained. The table highlights the distinction made above between traditional, so-called Ricardian gains, and two types of intra-industry Smithian gains from trade. The second row of the table identifies the different products that are traded, and also reflect the assumptions of the different theoretical models used to analyse these different types of trade. Traditional trade models assume homogenous goods, which leaves no scope for intra-industry trade. More advanced models of intra-industry trade assume heterogeneous goods. Horizontal intra-industry trade tends to be in final goods differentiated by brands or quality attributes, including the food niche products and some special high-value horticultural products. On the other hand, examples of vertical intra-industry trade are intermediate goods and outsourcing. Most trade models focus on countries as their unit of analysis, but in reality it is firms that are conducting the trade. Therefore it is important to identify the types of firms involved in these different types of trade. Ricardian trade can in theory involve any firms. However the traditional trade models are characterized by a couple of important assumptions: perfect competition and no economies of scale internal or external. The firms involved in horizontal intra-industry trade are not necessarily internationally integrated while the ones involved in vertical intra-industry trade will normally be firms linked to foreign firms through long-term affiliations or because they are part of global value chains (Humphrey and Schmitz). What is important from a theoretical point of view is that these firms are characterized by economies of scale and learning processes which introduces increasing returns and the possibility of specialization. The next row focuses on the relevant policy instruments. When considering Ricardian trade typically we focus on tariffs and tariffs-like barriers (e.g. quotas), these can include standards and regulations to the extent that these act as a tariff and therefore their tariff-equivalent can be calculated. But for intra-industry trade the importance of tariffs is smaller and the relevant policy instruments are standards, regulations, testing and conformity assessment. Due to their differentiated nature it gets crucial to be able to assess their characteristics. In certain cases these are information diffusion devices implicitly imposed within the production chain as in the case of global production chains (e.g. cars). From this it is clear why when analyzing standards and regulations we will consequently shift the focus of our attention to Smithian trade. However standards can also act as a barrier and be a tariff-like trade policy instrument as we have already discussed and this possibility is included in our theoretical framework when we are analyzing Ricardian type of trade. We also need to ask if and what externalities and market failures are associated with these kinds of trade. This is important to link the concept of Smithian trade gains to the relationships between standards and public goods. We argue that when considering traditional trade gains and Ricardian type of trade it is often assumed that there are no market failures nor externalities. On the opposite, when considering Smithian type of trade these are much more diffused. In particular horizontal intra-industry trade is clearly affected by information externalities (e.g. one potato affected by brown rot and all the shipment is destroyed), reputation mechanisms, learning effects (e.g. farmers can learn from their neighbour how to take care of certain pests for instance), etc. Similarly, a number of spillovers affect also vertical intra-industry trade, in particular the establishment of quality assurance systems, the fixed costs involved in setting up a system of standardisation and coordination among trading partners. Finally, what are the lessons from the point of view of needed public policies and collective action that we can derive from the distinction between Ricardian and Smithian type of trade? When considering Ricardian trade we are left with traditional public policy measures involving elimination of barriers or the so-called negative integration, the main achievement of the various GATTs rounds. Differently, when considering Smithian trade we need to consider a much wider and more complex set of policies, in particular when considering horizontal intra-industry trade the scope of public intervention is a consequence of the market failures and externalities described above. However we would argue that when dealing with vertical intra-industry trade the scope for public intervention should be maintained at a narrower level because we observe that private sector mechanisms of coordination tend to work well, surely there is scope for public policies to improve the business environment and facilitate trade but we would argue that the government shouldnt be involved in setting mechanisms for intra-firm coordination because it probably lack the right information and possibly also lack the right incentives. The bottom panel of the table also identifies the role regional agreements could play in trade facilitation. It becomes rapidly clear though that this can only be a minor part of the story. Table 2: DEEP INTEGRATION AND ROUTES TO GAINS FROM TRADE Traditional trade gains & trade policiesSmithian productivity enhancing gains from trade Horizontal IITVertical IIT TradeDriving forces (economic determinants)H-O comparative advantage: endowments, income, factors costsProductivity gains driven by product innovation and specialisation (including advertisement)Productivity gains driven by process innovation and thinner division of labour (slicing up production chain, e.g. Mex-US)Type of product affectedHomogeneous (either final or intermediate)Mostly final, differentiated by variety and by quality Includes more agricultural niche products. (process of de-commodification of commodities)Trade growth is in intermediate goods may be homogenous/interchangeable but quality needs monitoring and differentiation along production chain; mostly industrial but not only. Services outsourcingType of firms involved in this type of tradeAny firmmay be internationally integrated or subcontractingMore likely to be integrated; long term affiliation or subcontracting likelyRelevant Policy instrumentstariffs QRs etc. Any standards, regulations that are purely tariff-equivalents (especially time consuming controls done at port of import)Standards, regulations, conformity assessment Monitoring can be done on import but potentially very costlyStandards mostly Monitoring must be done along the whole prod chainMarket failures/ externalities associated with this kind of tradeAnyReputation, health, learning effects,lack of quality assurance systems, standardisation, general issues of business environmentAction needed to remove barriers: Public vs. PrivatePublic policy measures need to be addressedMore scope for public intervention here (especially in case of market failures: e.g. health issues that individual consumers cannot detect these interests may require public intervention i.e. driven by public welfare); NB avoid raising rivals costs standardsMuch scope for intra-firm resolution and private coordination; public policies needed for favourable environment; ROLE FOR RTASEliminate border barriers and equivalentquality standards ensuring protection of consumer and environment As above all quality and compatibility for processesgains from trade if RTA successful traditionalhigher profitability from niche products plus learning about value chain + economies of scopeeconomies of scale and technology transferImpact of trade expansion on povertyPoorest make less tradables so less able to profit from expansion of formal sector . But possible impact on employment and wages? Maybe hard for small firms to meet standards However there can be positive labour impact if larger can do soVery poor unlikely to be involved at all, but huge gains if somehow they can get into international value chains. However there can be positive labour impact (e.g. maquiladoras employment in Mexico) We see however an interesting paradox as we consider the various possible effects of deeper institutional integraton. Deeper institutional integration lowers transactions costs so that firms can more easily do business with others that they do not initially know. At the same time markets are clearly more integrated when they are less than totally anonymous, ie when participants have more information about each other, which is likely to increase trust. This implies that the kind of deep market integration promoted by institutions that cut transactions costs can create two forms of deep integration: a very comprehensive certification or standardisation regime which can permit trade between anonymous partners OR networks of repeat business This also raises the possibility that the nature of the relationship between firms is likely to be non-monotonic as deeper integration takes place. A first phase of deeper integration can lead to the possibility of investing in durable market relationships, the creation of quasi-integrated value chains; but that there could be a later step in the process where buyers and sellers can afford at very little cost to switch trading partners but still be assured of the requisite quality and reliability. Alternatively, it may be that purely market driven deep integration leads to one outcome whilst collective institutions are needed for the other. A paper by Leijonhufvud (2007) drawing on the unbundling concept surveyed by Baldwin (2006) suggests that value chains have some of the features of the intra-firm monopoly-monopsony elements that exist within firms for successive stages of the production process. Without referring to deep integration as such he suggests that the most important contribution institutions can make is to create thick markets where upstream and downstream partners have a free choice to switch contractors. In this scenario durable contracts and many potential partners could coexist as it would be the availability of alternatives for the other side rather than the opposite that induced each party to stay with the relationship. So, in looking at the depth of integration we need to consider the thickness as well as the depth of integration. The role for institutional deep integration may be most important in ensuring the correct balance. The Gains from Deep Integration In an early pioneering empirical analysis, and in the context of the Euro-Mediterranean free trade area, Hoekman and Konan (1999) concluded that: Despite the weakness of the datasets that are available, the major points that emerge from the analysis are fully consistent with the economic theory and analytical models. PTAs that are limited to the elimination of tariffs for merchandise trade flows are of limited value at best, and may as easily be welfare reducing as welfare enhancing. Similarly to the preceding discussion, they define deep integration as explicit actions by governments to reduce the market segmenting effect of domestic regulatory policies through coordination and cooperation. They go on to conclude that the earlier pessimistic conclusion may be overturned by appropriate forms of deep integration that eliminate regulatory barriers. But they are forced to make fairly arbitrary estimates of the impact on trade costs of integration. This raises the question of what are the possible sources of the gains from deep integration? The answer is multifaceted and complex, partly because deep integration involves a number of diverse areas as detailed above, and partly because there are typically spillover/externalities associated with aspects of deep integration. However, there are a number of possible gains that can be identified. Examples of the benefits of deep market from deep policy integration include: more niche market specialisation and the creation of more stable value chains. technology transfer and diffusion both through trade and FDI, pro-competitive gains from increasing import competition in an environment of imperfect competition, greater exploitation of economies of scale in production the greater use of intermediate inputs; the increased geographical dispersion of production through trade that supports the exploitation of different factor proportions for different parts of the production process local economies of scale through finer specialization and division of labour in production; externalities arising from institutional changes that lead to a wide increases in productivity. In addition to the potential, but once and for all, efficiency gains and losses, there may be welfare gains arising from growth effects induced by economic integration. There might be faster technical change and total factor productivity growth from the increased specialisation, and/or positive externalities between firms and/or sectors. These dynamic gains are typically more likely to arise in the presence of deep integration, and are often also associated with the presence of foreign direct investment In our work on EU Egypt we looked at the areas where the Association Agreement raise the possibility of genuine deep integration, bearing in mind that the impact of harmonisation will depend critically on the very nature of the deep integration process. Consider the barriers to deep integration identified in Table 1 above. Most of these are barriers to the contestability of domestic markets hence introducing a competition policy, introducing measures encouraging investment, reducing public procurement etc are all concerned with increasing the responsiveness, capacity, and attractiveness of the domestic market both for domestic and foreign firms (hence linking in to FDI). Introducing these elements of deep integration will therefore have a first order impact on the firms currently supplying the domestic market in the related and affected sectors. In the first instance this implies structural adjustment with winners and losers. In the second instance those processes of deeper integration are likely to increase the supply capacity and responsiveness within the domestic market (be it by domestic or foreign firms) driven, for example, by increased availability of credit for investment purposes, market expansion, productivity growth, improved supply (in terms of both price and quality) of intermediate inputs (e.g. telecommunications), and improved infrastructure. In contrast some of Table 1, is more directly related to improving the ability to access foreign markets. Here the impact is likely to be very different on those firms that are able to benefit from the enhanced market access as opposed to firms who must merely respond in the domestic market. Standards fall directly into this category. Consider, for example, the impact from the harmonization of standards: Exporting firms that can meet higher standards are likely to gain (or at least maintain) market access. In addition the ability to meet high standards may allows some (eg. developing country) exporters to become part of integrated value chains that divide up processes traditionally carried out in one production unit so they can be relocated anywhere in the world. Becoming part of such a vertically integrated supply chain may lead to increased exports, greater long term stability of exports which can lead to higher levels of investment, and might lead to positive productivity spillovers either directly via technology or via improved managerial practices and know-how. However, to set an entire domestic standards regime to assist just those firms exporting to, for example, the EU might be a small tail wagging a large dog. The new economic geography suggests that these relocations may not be uniform but frequently occur in clusters of activity, some of which may be associated with RTAs. The ability to take part in this disintegration and reallocation of the production chain depends on guaranteed output quality compatible with the requirements of the buyer and on the characteristics of the market. The introduction of standards means that domestic firms are likely to face higher costs and consumers are likely to pay higher prices. This may or may not be appropriate for any given domestic market. While facing higher domestic prices, domestic consumers are also likely to be getting better quality. In addition to the extent that standards generate public goods the costs will be lessened. It seems clear then that the importance of standards needs to be considered on a case by case basis. There is little point in worrying about the standards for some product which a given country scarcely produces. There is not a lot of point in worrying about applying, for example, EU standards for a product which a given country only produces for its domestic market. However, it is likely to be important to be concerned about standards issues for goods which the country produces and exports, or is capable of exporting in significant quantities, such as Egyptian potatoes for example. There is a very considerable private element in the demand and supply of standards and standards related services (see Holmes, Iacovone et al 2006) which can lead to market led deep integration. However, there is also the potential for market failure, arising from asymmetric information, externalities and spillovers which thus raises the potential need for policy intervention. That policy intervention can take the form of deeper institutional integration. For example, as we have shown above, a producer who wishes to comply with standards must purchase the standard, incur the compliance costs, incur regular and repeated testing and certification (the cost of this is likely to increase with the reputation of the certifier, which must itself incur accreditation costs). For the individual producer of goods or services, purchase of the standard is likely to be a small fraction of the marginal cost of compliance. But the fixed costs involved in setting up conformity assessment institutions and accreditation systems are too large for an individual producer, or even a small group of clients for a conformity assessment company, to sustain them. Despite the fact that the market can and does provide standards related activities, there are very extensive spill-over effects, whether literally through the environment or via learning or reputation effects, hence there is likely to be an under-utilisation of conformity assessment via market forces alone. Market failures could thus arise from any of the following: Asymmetric information: We cannot assume that with imperfect information about the benefits of standards that firms will privately purchase the right amount of standards and quality assurance etc. There is an informational public good that may need to be supplied. Learning Spillovers: Once firms do engage with international standards there is learning by doing from the adoption of standards by the firms that adopt them. This is not in itself a market failure provided that firms anticipate this: with imperfect information however the same asymmetric information issues arise. The experience gained can be transmitted to other firms however. Circulation of experience and learning form competitors and customers are well-documented ways in which technological spill-overs occur (von Hippel, 1986). It must not be assumed however that the transfer of knowledge is without cost. Learning by doing and learning by imitation involve increasing returns and falling but not zero marginal costs. Reputational spillovers: Probably one of the most important aspects of standards adoption is the market signalling aspect. For a large multi-product firm, investment in quality or brand image has spillover effects across its product range. This is less true for smaller firms. In the 1960s Made in Japan was a signal that signified low quality to many consumers: now it is the reverse. During their earlier periods of industrialisation a distinction between South Korea and Taiwan was that the former has a smaller number of large firms whose names were recognisable, while Taiwans engineering exports were largely from small subcontractors to US firms with no individual reputation. Environmental spillovers: Normally one firms adoption of a quality norm has only a small direct effect on any of its neighbours. However there may be cases where failure of one producer to confirm to standards actually can damage output from other firms. The most obvious case is bio- and environmental norms. For example if one agricultural producer allows chemical or biological contamination to take place, this can result in contamination of other producers. Where quality controls by importing countries are based on geographical regions the failure of one producer to respect quality norms can affect an entire crop. Economies of scale: We have stressed that for much standard related activity the private costs are very significant, but just as there are benefit spillovers there are also likely to be positive externalities on the cost side. The fact that we see the development of standards related activity in developed countries where the demand is greatest suggests that there is perhaps a critical mass below which an indigenous certification or accreditation system cannot be profitable: it is after all much cheaper per unit to inspect all the farms in a region than just one. Compatibility externalities: One of the most important aspects of technical norms is to ensure not merely that supposedly identical products really are so, but to ensure compatibility of products that have to be used with others. Within a value chain the issues are similar to those of ensuring common standards, but one important role of standards is to ensure for example that printers can be linked to any computer etc. Network externalities are an obvious example of this. The more people who use GSM phones in a country the more people existing users can potentially connect to. Much of this activity is carried out by private firms, e.g. standardising CD formats, but there is clearly room for public action in certain cases. As we noted above, Best (1990) argues that the American economy was transformed in the nineteenth century by the insistence of the army that all parts and component of rifles and other equipment be made interchangeable. Collective action and coordination failures and network externalities: The above is one special case of the general issue of coordination. There are various cases when market failures can arise because of lack of collective action and coordination among private individuals. In particular, whenever producers do perform joint marketing or the productions quality of one has an impact on the price and market access of the others (e.g. brown rot) the intervention of the government may be required to support or enforce the coordination and avoid free-riding. With regards to compliance with standards, if this is a costly process, individual firms will have the incentive to avoid compliance whenever they can rely on others compliance and escape the negative consequence of their non-compliance. However such a behaviour of one single producer may put at risk the entire sales of all the producers, therefore government oversight or regulations may be required to increase the costs of non-compliance and minimise free-riding. In certain cases, government intervention only needs to address asymmetric information failures to allow private self-enforcement to be re-established. At a more mundane level general acceptance of the compatibility principle can ensure that bicycle repair shops can stock standard tyres etc. Experience in East Asia suggests that many of these problems can be overcome and a high degree of market integration achieved by market forces alone. However as Suttons work shows many producers have difficulty breaking into the supply chains that emerge, though it is not obvious that this is itself a market failure as opposed to a lack of capability that needs to be addressed otherwise. In examining the scale of deep integration, in order to apply the test on standards upgrading versus importing the wrong norms, we need to examine whether the importing of norms is inherently capable of addressing externalities or other market failures. Such market failures could in principle be addressed through: A multilateral agreement, e.g. by ISO norms supported by the WTO A regional/bilateral agreement, e.g. RTA Autonomous government decisions of the individual countries The market, e.g. through FDI, detailed sub-contracting arrangements through the value chain, or by purely voluntary standards. Achieving deeper institutional integration The preceding sections detailed the different types of barriers to deep integration which may exist and which in their different ways make it more difficult for foreign firms to access domestic markets, and considered the possible welfare consequences of deeper integration. We have argued that regional integration is only one part of the question, but rightly or wrongly the European Union has embraced the idea that the regulatory integration that created market integration within Europe can be extended in its various trade agreements with the rest of the world. In this section, we consider the ways this can be addressed within the context of a regional trade agreement such as the Association Agreement, and the European Neighbourhood Policy Action Plan between Egypt and the EU. Barriers to foreign firms necessarily entail discrimination. The elimination of that discrimination can then be achieved in different ways and at different levels, and these are detailed in Table 3. In particular one can distinguish the notions of MFN, National Treatment, mutual recognition, and harmonisation. Most Favoured Nation (MFN) treatment: The MFN rule is the weakest form of deep integration and is an obligation to treat all foreign suppliers in the same way. It is rarely considered to be relevant in the context of regulations for goods, since it does not require markets to be opened to foreign competition but merely that however closed or open they are the same treatment is extended to all foreign suppliers, even if collectively they are not treated the same as home firms., But it is more complex for services. The WTO agreement on services, the General Agreement on Trade in Services GATS (see Box 1.1) requires MFN even for unscheduled and un-liberalised sectors or modes, so all foreign suppliers must be treated alike unless an MFN exemption has been agreed. This has never been a diplomatic issue but it is a constraint on domestic rule making: one cannot decide to admit service suppliers from the EU but not the US without either an MFN exemption or the negotiation of a fully fledged Regional Trade agreement compatible with GATS article V. But this provision is more like to be relevant as a constraint on partial liberalisation than for highly regulated markets. Table 3. Ways of Eliminating Barriers to Deeper Integration Type of barrier removalMeaningWTOGovernanceMost favoured nationDiscrimination against foreigners allowed, but not between them Key principle of GATT; GATS if there are no exemptionsMinimal constraints on autonomy, but says any liberalization must be non-discriminatoryNational treatmentMust treat foreign and home products equally once they are admitted Key principle of GATT, Article III GATS where commitments madeImpact may be slight without binding dispute settlement (DS). Could be substantial with binding DS if de facto implicit discrimination addressed. Sometimes impact underestimated Mutual recognitionEquivalence assumed, but can be subject to challenge SPS if equivalence can be shownDepends on freedom to choose partners and if there is binding DS.Harmonisation/ ApproximationEveryone has same/similar rulesTRIPs GATS Telecoms reference paper SPS & TBT some elementsExplicit agreement in detail needed in the EU qualified majority voting; at the WTO consensus if multilateral. But ISO & Codex can be decided by majority Approximation agreed in an RTA can reduce sovereignty  National Treatment: It was sometimes thought that the General Agreement on Tariffs and Trade (GATT) of 1947 only dealt with shallow integration but in fact Art .III calls for National Treatment and thus requires all domestic regulations and taxes to treat home produced and imported goods alike. Also the ban on quantitative restrictions in Article XI has been interpreted as covering bans on the sale of all goods of a particular type that obstruct imports even if they also affect home goods. The general obligation for National Treatment of GATT Article III and for those sectors so scheduled under GATS stipulates that taxes and regulations must be the same on imported as well as domestic goods, once the former have cleared customs. This obligation is weak if it is confined to saying that laws regulations and taxes cannot discriminate legally against foreign goods, but it is powerful if de facto discrimination is prohibited and there is binding dispute settlement (DS) system. The European Court of Justice has ruled that any differences between national regimes even if they are not discriminatory can be obstacles to trade and the country imposing them must demonstrate that the country imposing them has a solid public policy objective in mind that cannot be achieved another way. Therefore, though the term deep integration was coined in the 1990s to refer to trade agreements that incorporated integration processes that went beyond the obligations of the GATT itself. Within the EEC, the ECJ went further than the GATT but was building on it. It is rare that technical barriers to trade are written in such a way as to exclusively target imports, though it is possible for a technical regulation to fall more heavily on imports de jure. More common is the situation where regulations in one country are an implicit or de facto obstacle to imports because they are easier for local products to meet. Box 1 The General Agreement on Trade in Services (GATS) Part of the WTO agreement of 1994 What is covered? Each WTO member must list or schedule those sectors it is willing to open. These commitments are bound. Non scheduled sectors do not have to be opened up. But they subject to the MFN obligation unless the country files an exemption for them. There is a general but somewhat imprecise exemption for public services. What commitments apply to scheduled sectors? Countries listed how much and in what way they were willing to open each of 16 broad sectors with the possibility of excluding sub-sectors within them. For each open sector the country must schedule its willingness to open the sector for any or all of 4 modes: 1. Cross border supply of services 2. Supply of services to buyers who travel 3. Establishment of foreign suppliers 4. Cross border supply of services by temporary movement of people In each case the schedules distinguish between liberalisation (i.e. committing to allowing foreign firms to supply at all) and National Treatment i.e. letting them in on the same terms as locals. Many countries open their markets more than they are signed up for under GATS. These openings are reversible. In principle GATS affirms the right of countries to retain freedom for domestic regulation, though the border between this and trade is blurred. Services in RTAs: The rules on services in RTAs are separate from those on goods. The GATS agreement allows WTO members to sign preferential agreements on services with partners but only if substantially all services trade is covered. Some services RTAs however merely reaffirm GATS commitments.  Mutual Recognition and Harmonisation: The EU has been the prime example of deep integration and its regulations and jurisprudence have gone well beyond the idea of national treatment. The Court of Justice has declared that differences in rules can be seen as obstacles and have to be shown to be necessary. As a result the EU has long found it necessary harmonise rules. However this is a difficult political project even though the EU has a legislative apparatus. Hence the ECJ developed the notion of mutual recognition: products made to meet one member states norms must be accepted by others, unless they can show why they not acceptable. Unfortunately it did leave some scope for Member States to reject partner country goods if they fell below a certain quality threshold and it became necessary to move towards the new approach in which basic elements of product specifications were harmonised leaving freedom on inessentials and how to achieve aims. Mutual recognition would seem to be the ideal mechanism to ease trade frictions without reducing regulatory freedom. Sadly it is not so simple. The EU has been forced to agree basic standards and it has in fact found it very difficult to agree mutual recognition of testing and certification with even the US. Mutual recognition in practice requires some degree of approximation and a high degree of trust between parties. Hence it has not been widely used at the WTO or in RTAs. This brings us to the core issue of how far deep integration can be used to facilitate trade in RTAs, or put alternatively, to identify the mechanisms by which RTAs could attempt to address these barriers. Here there are actually two dimensions: (a) the extent to which the liberalization can and does discriminate between RTA and non RTA partners; and (b) which mechanism national treatment, mutual recognition, or harmonisation is used. These issues are summarized in Table 4. With regard to the issue of discrimination between RTA and non-RTA partner, in general an agreement to liberalise in a RTA can be preferential or MFN. The latter implies that any advantages given to the RTA partner country are then extended to others for example: by rules e.g. the TRIPS agreement says that any increased protection of Intellectual property rights must be extended on an MFN basis to all partners; so US Jordan FTA extends IPR protection to all WTO members. Some RTAs contain provisions that liberalisation in subsequent RTAs must be extended to signatories of earlier deals if not all states de facto when it is very hard, if not impossible, to exclude others from benefits. For example with simplified bureaucratic procedures it is in many cases hard to apply these to some importers / exporters and not others. Similarly, agreeing to use EU standards (or ISO standards) would benefit all who sell into EU, and/or who sell into the domestic market. Introducing or changing a competition policy applies to all firms in the market. Investors can buy into a firm that is based in home that has benefited from a relaxation of investment rules. In considering standards, it appears that if deep integration is to be effective it must be thorough-going and contain some form of harmonisation. A major practical problem is that the EU calls on its partners to approximate their legislation to that of the EU and with mutual recognition of conformity assessment very much as a final step. Even the recent generation of agreements contain largely aspiration and only the mildest of commitments. By contrast very few other RTAs such as NAFTA go beyond the National Treatment concept and hence are unlikely to be able to deliver much more than the WTO. Other RTAs such a Mercosur have lofty ambitions for harmonisation but have de facto made little progress. This applies to most areas of regulatory harmonisation (Bourgeois et. al., 2007). In fact, as discussed earlier, a very considerable amount of deep integration can be done by the private sector. Standards are often private and conformity assessment is typically a private activity. Nevertheless it rests on a public infrastructure and there is little getting away from the fact that deep integration cannot confer real market access gains without loss of sovereignty. The EUs relations with its immediate EEA neighbours exemplify forced harmonisation with the bare minimum of consultation. These countries, like candidates, however have little option however but to align their economies with their dominant market. For the ENP partners the degree of market integration is not so close so that real choices have to be made. There is increasing pressure to create some for of common market place in which all forms of regulatory conditions are approximated. Standards and technical regulations are the most obvious elements in this process but as the EU itself has found, effort is needed to remove a range of other regulatory barriers such as restrictions on government procurement or competition laws that allow local firms to create entry barriers. We would argue that in these regulatory domains it is less likely that there will be adverse discrimination against 3rd parties. Putting it simply, adapting non-trade rules to make life easier for EU firms usually also makes it easier for any firm able to sell into the EU. Table 4. Ways in which Regulatory Barriers could be dealt with in an RTA PREFERENTIALMFN NATIONAL TREATMENT*MUTUAL RECOG.*HARMON-ISATIONStandards: (SPS, TBT)MR with some partnersAgreement that both parties will adopt ISO etc normsGenerally required by GATT Art III anywayFor conformity assessment likely to be preferential EU presses for this, but limited value without CA MR InvestmentDiscrimination possible in industry so long as trade not affected. But de factor hard to apply if 3rd country can buy into the industryGATS require- meant unless Art V satisfied. Agree to apply same rules to foreign firms as homeMost likely to apply for servicesIPR Must be MFN under TRIPS Rare case where you have to extend Generally required under TRIPSunlikelyEU & US call for this in some areasTrade defenceEEA removes ADUnusualn/aServicesMust satisfy GATS V. Preferences hard to apply if 3rd country can buy into firms Quite likely; GATS obliges MFN for all sectors even if not scheduled unless exemption taken out in 1994Unlike goods only required for scheduled sectors; Preferences possible But subject to Art V Home country regulation Possible, but GATS has rules.PossibleGovernment procurementCan be donePP not covered by GATT, only GPA PossibleWould apply to approved listsCompetition PolicyUnlikely but could apply to takeover rulesMost likelywould affect all partnersPossible to apply to some or all Would imply MR of decisionsPossible State aids rulesHard to do except for CVDs.State aid controls would affect all partners??EU seeks to secure: bonus removal of CVD rules Conclusion Our concern in this paper has been to explore the ways in which the new types of market integration differ from earlier trade patterns. Global integration is proceeding apace, but unevenly. We can speak of deep integration taking a number of forms. The removal of trade barriers is more and more largely focussed on domestic regulations while trade flows increasingly concern processes or tasks, including services rather than finished products. A key criterion for the efficiency or desirability of trade agreements is the extent to which they are actually necessary to facilitate this process. When tariff barriers are low the most important contribution trade deals can make to facilitate trade is the removal of NTBs and the harmonisation of regulations. The EU is embracing this philosophy in its new generation of FTAs including, where it can in its EPAs. Limited attention has been paid to the trade-offs involved in the process. Whilst classic trade liberalisation can lead to trade diversion, and as we see above there is a need to ensure that harmonisation within regions of PTAs does not exclude others the new forms of integration can also lead to exclusion of suppliers from countries where there is a general structural inability to meet standards. No amount of price cutting will induce buyers to accept products that are below quality minima. At the same time within countries, as goods become more differentiated by supplier, there will be producers who can and cannot meet international demand. Clusters of economic activity are developing that are not necessarily defined by geographical boundaries, although geography still does matter a lot. We also see business networks being established within firms but more interestingly through sub-contracting and outsourcing. These links are very much market driven, rather than by public policy. They also seem to relate to narrower elements of comparative advantage than used to be thought. Comparative advantage may be identified for a particular processes or task for a specific product, not for a whole industry. The result is that we can secure major new sources of gain but ones that are potentially unevenly distributed, horizontally and vertically. Horizontally the creation of tougher and more demanding standards is likely to create inequalities between producers who can and cannot meet the quality demands. This distinction may be as much within countries as between them, though new EU rules on food safety require make demands of the national; infrastructure. There is thus scope for multiple types of market failure, in terms of spillover effects, lack of public goods etc. The challenge is clear but policy response is not necessarily obvious however: aiding small producers to try and meet standards that are always going to be impossible is futile. Vertically the setting of standards by private firms or consortia raises questions about the distribution of rents from innovation. The classic debate about the role of supermarkets in the farming chain highlights this. In all cases the question arises: is there governance challenge or will the market deliver the least bad deep integration . And if there is a governance gap at what level should it be addressed - by individual governments deciding how far domestic regulation should be driven by market access considerations? In regional or bilateral agreements or multilaterally? References Baldwin, R. (2006) "Globalisation: The great unbundling(s)", report for the Finnish Prime Ministers Office (in context of EU Presidency 2006). Best, M. H. (1990), The New Competition Polity Press. Bourgeois,Jacques, Kamala Dawar, and Simon J. Evenett. (2007) "A Comparative Analysis of Selected Provisions In Free Trade Agreements." Report prepared for DG TRADE, European Commission El Garf, Mona, Michael Gasiorek, Ahmed F. Ghoneim (Cairo University), Peter Holmes et al (2007) Examining the Deep Integration Aspects of the EU-South Mediterranean Countries: Comparing the Barcelona Process and Neighbourhood Policy, the Case of Egypt Study for FEMISE programme Helliwell, J.F (1998) How Much Do National Borders Matter? Brookings Helpman, E., M.J. Melitz and S.R. Yeaple (2004). Export versus FDI, American Economic Review. 94(1) Hoekman B. and L. Alan Winters Multilateralizing Deep Regional Integration: A Developing Country Perspective www.wto.org/english/tratop_e/region_e/conference_sept07_e.htm Hoekman, Bernard and Konan, Denise Eby, "Deep Integration, Nondiscrimination, and Euro-Mediterranean Free Trade" (May 1999). World Bank Policy Research Working Paper No. 2130. Available at SSRN:  HYPERLINK "http://ssrn.com/abstract=604992" http://ssrn.com/abstract=604992 Holmes, P and Iacovone L et al (2006) Capacity-Building to Meet International Standards as Public Goods, UNIDO  HYPERLINK "http://www.unido.org/file-storage/download?file_id=60028" www.unido.org/file-storage/download?file_id=60028 Humphrey, J. (2008) Private Standards, Small Farmers and Donor Policy: EUREPGAP in Kenya IDS forthcoming0 Humphrey, J. and Schmitz, H. (2002), Developing Countries Firms in the World Economy: Governance and Upgrading in Global Value Chains, Institut fr Entwicklung und Frieden, Duisburg, Report 61. Jaffee, S. and Henson, S. (2004), Standards and Agro-food Exports from Developing Countries: rebalancing the debate, World Bank Policy Research Working Paper No. 3348. Lawrence R.Z. (1996). Regionalism, Multilateralism and Deeper Integration, Brookings. Leijonhufvud, Axel (2007) "The Individual, the Market and the Division of Labor in Society," Capitalism and Society: Vol. 2 : Iss. 2, Article 3. Available at: http://www.bepress.com/cas/vol2/iss2/art3 Mattoo, Aaditya and Chen, Maggie Xiaoyang, 2004. " HYPERLINK "http://ideas.repec.org/p/wbk/wbrwps/3458.html" Regionalism in standards - good or bad for trade?,"  HYPERLINK "http://ideas.repec.org/s/wbk/wbrwps.html" Policy Research Working Paper Series 3458 Melitz, M.J. (2003) The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity Econometrica 71 (6) , 16951725 Sutton, J. (n.d.), The Auto-component Supply Chain in China and India: A Benchmarking Study. Available online at: http://sticerd.lse.ac.uk/dps/ei/ei34.pdf Von Hippel, E. (1986), Lead Users: A Source of Novel Product Concepts, Management Science, 32, 7 pp. 791-805. Williamson, O. E. (1975), Markets and Hierarchies, (New York: Free Press).  We could also have used the term Heckscher-Ohlin trade.  A number of recent theoretical models have shifted their attention from countries to firms (cf. Melitz, 2003; Melitz, Helpman and Yeaple, 2004)  We have drawn on this paper here.  See Hoekman and Winters 2007  See Mattoo and Chen 2004  See Humphrey forthcoming     PAGE 1 PAGE 23 =>]5 8 B C Z [ \ ] ^ _ ɸ𝈝tbI5'hCNhCN0J(OJQJ^JmHnHu0jhCNhCN0J(OJQJU^JmHnHu#hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u(jhCNhCN5CJOJQJU^JhCNhCN5CJOJQJ^JhCNhCNCJaJ hCNhCNCJ OJQJ^JaJ hCNhCNCJOJQJ^JhCNhCNCJ$OJQJ^JhCNhCNCJ$hCNhCNCJ hCNhCN=>O\]5 6 7 8 B $@&^a$$@&a$$@&a$$0@&^`0a$ H B \ ]  p G '(1F$EƀF.a$$a$$@&a$, ! , ! _ { | } ~ ٹٌx`xF`x`2j}hCNhCNOJQJU^JmHnHu/jhCNhCNOJQJU^JmHnHu&hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u0jhCNhCN0J(OJQJU^JmHnHu>jhCNhCN>*B*OJQJU^JmHnHphu'hCNhCN0J(OJQJ^JmHnHu#hCNhCNOJQJ^JmHnHu         1 2 dzsYssǟdzG#hCNhCNOJQJ^JmHnHu2jwhCNhCNOJQJU^JmHnHu/jhCNhCNOJQJU^JmHnHu&hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u'hCNhCN0J(OJQJ^JmHnHu0jhCNhCN0J(OJQJU^JmHnHu>jhCNhCN>*B*OJQJU^JmHnHphu2 3 4 6 7 N O P j k l m n o p q r dzsYssǟdzG#hCNhCNOJQJ^JmHnHu2jqhCNhCNOJQJU^JmHnHu/jhCNhCNOJQJU^JmHnHu&hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u'hCNhCN0J(OJQJ^JmHnHu0jhCNhCN0J(OJQJU^JmHnHu>jhCNhCN>*B*OJQJU^JmHnHphu dzsYssǟdzG#hCNhCNOJQJ^JmHnHu2jkhCNhCNOJQJU^JmHnHu/jhCNhCNOJQJU^JmHnHu&hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u'hCNhCN0J(OJQJ^JmHnHu0jhCNhCN0J(OJQJU^JmHnHu>jhCNhCN>*B*OJQJU^JmHnHphu $ % & @ A B D E F G H I e f dzsYssǟdzG#hCNhCNOJQJ^JmHnHu2jehCNhCNOJQJU^JmHnHu/jhCNhCNOJQJU^JmHnHu&hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u'hCNhCN0J(OJQJ^JmHnHu0jhCNhCN0J(OJQJU^JmHnHu>jhCNhCN>*B*OJQJU^JmHnHphuf g h j k u v w dzsYssǟD:1hCNhCNCJhCNhCN5CJ(jhCNhCN5CJOJQJU^J2j_hCNhCNOJQJU^JmHnHu/jhCNhCNOJQJU^JmHnHu&hCNhCNOJQJ^JmHnHu'hCNhCNOJQJ^JmHnHsH u'hCNhCN0J(OJQJ^JmHnHu0jhCNhCN0J(OJQJU^JmHnHu>jhCNhCN>*B*OJQJU^JmHnHphu ]h>"?"I"Q"7788j8k8|899f9q9999:v:: ;;G;d;;;;<o=G>>i?5@6@@@BBkB0Q1Qoo hCNhIVjhCNhIV0JUhCNhCN56]hCNhCN5OJQJhCNhCN5\]hCNhCN5CJhCNhCN0JCJnH tH hCNhCN0J5CJnH tH hCNhCN5hCNhCN6 hCNhCN3mnm n $$'(%,&,wC1Eƀl&CEƀF.1 &,..447777788akd$$IflFprt"p       t0t"    4 lal $$$$Ifa$$$$a$1 88878F8j8k8|8Ekd$$Iflrprt"p          t0t"4 lal $$$$Ifa$|888899909Ekd$$$Iflrprt"p          t0t"4 lal $$$$Ifa$09E9J9e9f9q9u9y9Ekds $$Iflrprt"p     t0t"4 lal $$$$Ifa$y9999999E5$$$ \$Ifa$kd $$Iflrprt"p     t0t"4 lal $$$$Ifa$99999::#:Ekd $$Iflrprt"p     t0t"4 lal $$$$Ifa$#:3:7:u:v::::Ekd $$Iflrprt"p     t0t"4 lal $$$$Ifa$:: ; ;;;;0;Ekd?$$Iflrprt"p     t0t"4 lal $$$$Ifa$0;F;G;d;;;;Ekdr$$Iflrprt"p     t0t"4 lal $$$$Ifa$;;;;;;;dXXXXX $$$$Ifa$kd$$Ifl\prt"p      t0t"4 lal;;;;<QOOO1kd$$Iflrprt"p     t0t"4 lal<<H>YR"$ & F 8hEƀF.^a$R"$ & F 8EƀF.^a$H>{>>YR"$ & F 8hEƀF.^a$R"$ & F 8hEƀF.^a$>i?@YR"$ & F 8hEƀF.^a$R"$ & F 8hEƀF.^a$@5ABYR"$ & F 8hEƀF.^a$R"$ & F 8hEƀF.^a$BkBCCCD^H_HJJP_F$EƀF.a$$a$1R"$ & F 8hEƀF .^a$ PPWW)[*[ ^ ^__bbggjjllFoGossRwSwa{b{56/)1/0ށ̓ĄOW^jkvׇVWXY[iՉ /;ŊmۋμyhCNhIVB*CJph"hCNhIV56CJaJmH sH hCNhIVCJhCNhIV56CJaJhCNhIVB*CJph#hCNhIV56B*CJaJph hCNhIVhCNhIV56CJhCNhIV56hCNhIV5CJaJhCNhIV5./0Yekd$$Ifl4JF< #`t`p00(    4 la4f4$If $If` ށybYYYY $$Ifa$kd$$Ifl4\< # 0 0 0(4 la4f49pq΃g^^^Q^^^ $$If`a$ $$Ifa$kd$$Ifl\< #tp 0(4 la4Ą̈́Ng^Q^^ $$If`a$ $$Ifa$kd$$Ifl\< #tp 0(4 la4NOk%&ctug^^^^^^^^ $$Ifa$kdO$$Ifl\< #tp 0(4 la4 gZPPP $If` $$If`a$kd$$Ifl\< #tp 0(4 la45,,&$If $$Ifa$kd$$IflF\< #tp  (0(4 la4p(uvׇTUaXN $If` $$Ifa$kd+$$Ifl\< #tp 0(4 la4$IfUVWXYZgZPPP $If` $$If`a$kd$$Ifl\< #tp 0(4 la4Z[iՉ7.(($If $$Ifa$kd$$Ifl\< #tp  (0(4 la4p(Չ /;ĊaX $$Ifa$kd$$Ifl\< #tp 0(4 la4$IfĊŊmۋg^XXX$If $$Ifa$kd$$Ifl\< #tp 0(4 la4gecc1)kd$$Ifl\< #tp 0(4 la4TWnu{|klߵǹȹɹ~-L5"#+2ohCNhIV6CJnHtHhCNhIVCJnHtHhCNhIVnHtHhCNhIV5CJnHtHhCNhIV5jh+0JUhCNhIV6hIV hCNhIVhCNhIVmH sH hCNhIVB*mH phsH 7sw•SQQQ1Q1 & F h88xEƀF^81x^Q1 & F h88xEƀF^8•uvϗԛ^S$ & FEƀF^`a$$a$1F$EƀF.a$ԛyWS$ & FEƀF^`a$S$ & FEƀF^`a$yۜWS$ & FEƀF^`a$S$ & FEƀF^`a$ۜWS$ & FEƀF^`a$S$ & FEƀF^`a$HIGHIJMN1S$ & FEƀF^`a$ȹɹ~./#$1$a$C1Eƀ&.WiI"$ & F EƀFa$I"$ & F EƀFa$1Wkf"$a$I"$ & F EƀFa$I"$ & F EƀFa$>?WX12o $$$$Ifa$ $$$7$8$H$a$1$a$F$EƀF.a$ -l`````` $$$$Ifa$kd$$If<\x4!t  p     04 <al`````````` $$$$Ifa$kd$$If<\x4!t  p    04 <a !Al```` $$$$Ifa$kd!$$If<\x4!t p  04 <a[d\ 8@E)\^DE˹ˤ˗ˎˆ˂xhCNhIV56hIVhCNhIVH*hCNhIV0J3jhCNhIV0JUhCNhIVCJhCNhIV56CJhCNhIV5CJhCNhIV6 hCNhIVhCNhIV5hCNhIVCJnHtHhCNhIV5CJnHtHhCNhIVnHtH,,pl`````````` $$$$Ifa$kd."$$If<\x4!t p  04 <a `[YYYYY1$a$kdN#$$If<\x4!t p  04 <a $$$$Ifa$ Z[\ 8@'( $$Ifa$1KIDI?I$a$$a$1lkd|$$$IflY""  t 04 lap G1$EƀPIfiI$ & FEƀFa$I$ & FEƀFa$178OP89EFSXl{ $$$$Ifa$$ & Fa$1$a$I$ & FEƀFa$E(6R[$7an  sg0࿴tjthCNhCN0J4\hCNhCNCJaJhCNhCN\hCNhCNPJ_H nHtHhCNhCN5\ hCNhCNjhCNhIV0JUhCNhIVmH sH  hCNhIVhCNhIV0JCJH*nH tH hCNhIV5CJhCNhIV0JCJnH tH hCNhIV0J5CJnH tH )KC77777 $$$$Ifa$kd%$$Iflֈ\ D% T t04 laK#A5kd &$$Ifl3ֈ\ D% T t04 la $$$$Ifa$ACvJkd'$$Iflr\ D% @ t04 la $$$$Ifa$'(67kd'$$Iflֈ\ D% T t04 la $$$$Ifa$6EMQR[pq]kd($$Ifl\\ D% H t04 la $$$$Ifa$ H^mu $$$$Ifa$"C77777 $$$$Ifa$kd)$$Iflֈ\ D% T t04 la"#$7b7kd*$$Iflֈ\ D% T t04 la $$$$Ifa$+7kdo+$$Iflֈ\ D% T t04 la $$$$Ifa$+-/`ab72$a$kd_,$$Iflֈ\ D% T t04 la $$$$Ifa$bcn./] ^ $%fgtug $7$8$H$a$F$EƀF.a$1gh01PQz/$-`-a$ $7$8$H$a$$a$/$a$fg~ê}aD8hCNhCN0J(>*B*CJOJQJ^JaJmH phsH 7jO-hCNhCNCJOJQJU^JaJmH sH .hCNhCN5CJOJQJ\^JaJmH sH (hCNhCNCJOJQJ^JaJmH sH 1jhCNhCNCJOJQJU^JaJmH sH hCNhCNPJ_H nHtH$hCNhCN0J(>*B*CJaJphhCNhCNCJaJjhCNhCNCJUaJPvRSzêshhYhhhYhUKGKGh o?jh o?0JUhCNhCNhCN0J(>*B*phjhCNhCNUhCNhCN0J56]hCNhCNPJ_H nHtH hCNhCN(hCNhCNCJOJQJ^JaJmH sH 1jhCNhCNCJOJQJU^JaJmH sH 8hCNhCN0J(>*B*CJOJQJ^JaJmH phsH >hCNhCN0J(5>*B*CJOJQJ\^JaJmH phsH z{   " # % & ( ) + , 7 &`#$ $a$gd+ $a$gdIV $7$8$H$a$gd o? $7$8$H$a$    ! # $ & ' ) * , - 3 4 5 6 7 9 : @ A C D E G H I ͼͱ߭hCNh+0JmHnHuh o?0JmHnHu h o?0Jjh o?0JUjh=n!Uh=n!h o?jh o?0JUh+jh+0JU!7 8 9 E F G H I  $7$8$H$a$gd o?&`#$h]h4&P 1hBP. A!"#$% }DyK _Toc197161446}DyK _Toc197161446}DyK _Toc197161447}DyK _Toc197161447}DyK _Toc197161448}DyK _Toc197161448}DyK _Toc197161469}DyK _Toc197161469}DyK _Toc197161470}DyK _Toc197161470}DyK _Toc197161471}DyK _Toc197161471$$Ifl!vh5p5 5 #vp#v :V l t0t"5p5 /  /  / /  / / 4alM$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  /  / / /  /  / /  / /  4alM$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  /  / /  / /  /  /  / / 4al1$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / 4al1$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / 4al1$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / 4al1$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / 4al1$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / 4al1$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / 4al $$Ifl!vh5p5 55#vp#v #v:V l t0t"5p5 5/  /  /  / /  / 4al?$$Ifl!vh5p5555#vp#v:V l t0t"5p5/  /  / / /  /  / /  / / 4al$$If4!vh5t5p50#vt#vp#v0:V l4J0(++5t5p50/ / 4a4f4$$If4!vh5t5p5 5#vt#vp#v #v:V l40(++505 5/ / 4a4f4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 54a4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 54a4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 54a4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 5/ 4a4:$$If4!vh5t5p5 5#vt#vp#v #v:V lF (0(,5t5p5 5/ 4a4p($$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 5/ 4a4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 5/ 4a46$$If4!vh5t5p5 5#vt#vp#v #v:V l (0(,5t5p5 5/ 4a4p($$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 5/ 4a4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 5/ 4a4$$If4!vh5t5p5 5#vt#vp#v #v:V l0(,5t5p5 5/ 4a4:$$If!vh5t5p55 #vt#vp#v#v :V <05t5p55 /  /  /  /  / / / /  44 <a:$$If!vh5t5p55 #vt#vp#v#v :V <05t5p55 /  /  /  /  / /  / / 44 <a$$If!vh5t5p55 #vt#vp#v#v :V <05t5p55 /  /  /  / / / 44 <a$$If!vh5t5p55 #vt#vp#v#v :V <05t5p55 /  /  /  / / / 44 <a,$$If!vh5t5p55 #vt#vp#v#v :V <05t5p55 /  /  /  / / / / 44 <a$$If!vh5"#v":V l  t 05"4p $$If!vh5 55555T#v #v#v#v#v#vT:V l t05 55555T4$$If!vh5 55555T#v #v#v#v#v#vT:V l3 t05 55555T4$$If!vh5 5555@ #v #v#v#v#v@ :V l t05 5555@ 4$$If!vh5 55555T#v #v#v#v#v#vT:V l t05 55555T4$$If!vh5 555H#v #v#v#vH:V l t05 555H4$$If!vh5 55555T#v #v#v#v#v#vT:V l t05 55555T4$$If!vh5 55555T#v #v#v#v#v#vT:V l t05 55555T4$$If!vh5 55555T#v #v#v#v#v#vT:V l t05 55555T4$$If!vh5 55555T#v #v#v#v#v#vT:V l t05 55555T4KDyK 2www.unido.org/file-storage/download?file_id=60028yK rhttp://www.unido.org/file-storage/download?file_id=600286F@F Normal$a$CJ_HaJmH sH tH b@b  Heading 1$ & F <@&5CJ KH OJQJ\^JaJ d@d  Heading 2$ & F<@& 56CJOJQJ\]^JaJ^@^  Heading 3$ & F<@&5CJOJQJ\^JaJR@R  Heading 4$ & F<@&5CJ\aJT@T  Heading 5 & F<@&56CJ\]aJN@N  Heading 6 & F<@&5CJ\aJ@@@  Heading 7 & F<@&F@F  Heading 8 & F<@&6]T @T  Heading 9 & F<@&CJOJQJ^JaJDA@D Default Paragraph FontVi@V  Table Normal :V 44 la (k(No List pOp Heading 1 Char Char15CJ KH OJQJ\^J_HaJ mH sH tH udOd Char Char19356CJOJQJ\]^J_HaJmH sH tH u^O^ Char Char18-5CJOJQJ\^J_HaJmH sH tH uX@"X  Footnote Text$A$]^a$CJaJ@&@1@ Footnote ReferenceH*FOAF Body Text Char1\^JaJtH ^0@R^ List Bullet" xx^` mH sH tH 6B@b6 2 Body TextmH sH LOqL Char Char6CJ\_HaJmH sH tH u82@8 List 2^`>6@>  List Bullet 2  & FHC@H Body Text Indentx^\N@\ Body Text First Indent 2h^h`JOJ Char Char3CJ_HaJmH sH tH uV @V Footer$ !dh7$8$a$mH sH tH u.)@. Page NumbernOn CaptionCaption + leftx^CJPJaJmH nHsH tH8"@8 Caption 5CJ\aJJOJ Body Text 8! & FxxmH sH DOa"D nbodytext"<<`5\FOa2F lbodytext#<<^`B'AB Comment ReferenceCJaJ<R<  Comment Text%CJaJ@jQR@ Comment Subject&5\H@rH  Balloon Text'CJOJQJ^JaJ4U@4 Hyperlink >*phfM@af Body Text First Indent)$x`a$ PJnHtHN@N Header*$ 9r a$PJmH nHsH tHZOZ MTDisplayEquation+ l ^mH sH &@& TOC 1,.. TOC 2 -^.. TOC 3 .^<+@<  Endnote Text/CJaJ>*> Endnote ReferenceH*LOaL 3Style Body Text + Left1aJLO!L Body Text CharCJ_HaJmH sH tH NO"1N 1Style Body Text + Left CharOA a.X@Q. Emphasis6]0Ig{I>/JMI=>O78B'( mnmn %$&$&&,,/////000070F0j0k0|000011101E1J1e1f1q1u1y1111111111122#23272u2v22222 3 333303F3G3d333333333333^@_@BBHHOO)S*S V VWWZZ__byyyyzz||N}O}~~~~uvUVZ[ ĂłswuvϏԓy۔HIGHMN'(QR#$`aJ @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0@0 @ 0@0@0@0@0@0@0@0@10@10@10@10@10@10@0@0@0@0@0@0@0@0@0@0@0@0@0@0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0@0@0@0@0@0 @0@0@0@0@0@0 @0@0@0@0@0@0 @0@0@0@0@0@0 @0@0@0@0@0@0@0@0 @0@0@0@0@0@0 @0@0@0@0@0@0 @0@0@0@0@0@0@0@0@0@0@0 @0@0@0@10@10@10@10@10@10@10@10@10@10@10@10@10@10@10@10@10K00`10K004rK0010K00c0K00c0K00d0K0 0!0?r@0 K0@0ALr@0 K0B0Cr@0 K0D0Er@0 K0F0Gr@0 K0H0I,r@0 0=>O\]5678B\]pG'( mnmn %$&$&&,,/////000070F0j0k0|000011101E1J1e1f1q1u1y1111111111122#23272u2v22222 3 333303F3G3d333333333333344H6{66i7859:k:;;;<^@_@BBHHOO)S*S V VWWZZ__bbddFgGgkkRoSoasbs5x6xx/y0yYyyyyyyyyyyzyzzz{9{p{q{{|||||}N}O}k}}%~&~c~t~u~~~~~~~~~~~uvTUVWXYZ[iՁ /;ĂłmۃswuvϏԓy۔HIGHIJMNȱɱ~./#$.W>?WX12o -!A,pZ[\ 8@'(78OP89EFSXl{K#ACv'(6EMQR[pqH^mu"#$7b+-/`abcn./]^$%f g    t u   g h     01PQz{ "%(+789EFGJ000000000000000000000000,0808,08,08,08,08,08,0800 010101010101010101010101010 010n10n10n10n10n10n10n10n10n10n10n10n0n0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n0n0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 0n 10n10n10n "0n "0n "0n "0n "0n "0n "0n "0n "0n10n0n10n 010;10;10;10;10;10;10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@10;@)0;@0; @0; @0;@0; @0; `0; `0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0;@0;@0; @0;@0; @0; @0; @0; @0; @0; @0; @0; @0; @0;@0;@0; @0;@0;@0; @0; @0; @0; @0; @0; @0; @0;@0; @0; @0; @0; @0; @0; @0; @0; @0;@0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @0; @)0;@10;@10; 10;@10; 10;@10;@10;@10; 010@10@0@10@10@10 0 0 0 0 0 0 0 0@10@10@10@10@10@10@10@10@10@10@10@10@1010@0@10@10@10@10@10@10@10@0@10@10@10@0@10@10@10@10@10@10@10@10@10@10 "0 "0 "0 "0@"0 0@0@10@10@10@10@10@10@0@0 @0 @0 @0 @0 @0 @0 @0@0@0 @0 @0 @0 @0@0 @0@0@0 @0@0@0@0 @0 @0 @0 @0 @0 @0 @0@0 @0 @0@0@0@0 @0@0@0@0@0@0 @0 @0@10@10@10@10@10@10@10@10@10@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@10 @0 @10@0@10@0@10@10@10@10 0 0 0@0@10@10@10@10@10(0@10@0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0@0@0 @0 @0 @0 @0 @0 @0@0@0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0@0@0@0 @0 @0@0@0@0 @0@0@0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0 @0@10 0@10c@10c@10c@10c@10c@10c@10c10c10c10c10c10c10c0c0c0c0c0c0c0c/0c/0c0c0c0c0c/0c/0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c0c@0@0@0@0@0@0 00&@0@0@0@0@0@0@0@0@0@0 00=>O\]5678B\]pG'( mnmn %$&$&&,,/////000070F0j0k0|000011101E1J1e1f1q1u1y1111111111122#23272u2v22222 3 333303F3G3d333333333333344H6{66i7859:k:;;;<^@_@BBHHOO)S*S V VWWZZ__bbddFgGgkkRoSoasbs5x6xx/y0yYyyyyyyyyyyzyzzz{9{p{q{{|||||}N}O}k}}%~&~c~t~u~~~~~~~~~~~uvTUVWXYZ[iՁ /;ĂłmۃswuvϏԓy۔HIGHIJMNȱɱ~./#$.W>WX12o -!A,pZ[\ 8@'(78OP89EFSXl{K#ACv'(6EMQR[pqH^mu"#$7b+-/`abcnt  g   PQz #&),J000000000000000000000000,0808,08,08,08,08,08,0800 010101010101010101010101010 010n10n10n10n10n10n10n10n10n10n10n10n0n0n 0n 0n 0n0n 0n 0n 0n 0n 0n0n 0n 0n 0n 0n 0n0n 0n 0n 0n 0n 0n0n 0n 0n 0n 0n 0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n0n 0n 0n 0n 0n< 0n0n0n0n0n0n10n10n10n "0n "0n "0n "0n "0n "0n "0n "0n "0n10n0n10n 010;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;10;)0;0; 0; 0;0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0;0;0; 0;0; 0; 0; 0; 0; 0; 0; 0; 0; 0;0;0; 0;0;0; 0; 0; 0; 0; 0; 0; 0;0; 0; 0; 0; 0; 0; 0; 0; 0;0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0; 0;0;0; 0; )0;10;10; 10;10; 10;10;10;10; 0@10@@10@@0@@10@@10@@10@@ 0@@ 0@@ 0@@ 0@@ 0@@ 0@@ 0@@ 0@@10@@10@@10@@10@@10@@10@@10@@10@@10@@10@@10@@10@@10@@10@@0@@10@@10@@10@@10@@10@@10@@10@@0@@10@@10@@10@@0@@10@@10@@10@@10@@10@@10@@10@@10@@10@@10@@ "0@@ "0@@ "0@@ "0@@"0@@ 0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@0@ 0@ 0@ 0@0@0@0@0@0@0*@0@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@0t@ 0H00H00H0&0 H00H00H00@00%H0&0H00H0 0 H0%0H0 0H00H00H00H00H00H00H00H0)0K00I0I00I00I00I00I00 '''*_ 2 f EI B &,8|809y99#::0;;;<H>>@BP/NUZՉĊ•ԛyۜWKA6"+bgz7 I H BZ]}3Okmnp%ADEGgvfRI X%ĕX%ĕX%ĕX%ĕX%ĕX%ĕ̕XXXX  #*!!l,2$[ܧ2W_(Pנ@0(  B S  ?I _Toc197160611 _Toc197161399 _Toc197161441 _Toc197160612 _Toc197161400 _Toc197161442 _Toc197160613 _Toc197161401 _Toc197161443 _Toc197160614 _Toc197161402 _Toc197161444 _Toc197160615 _Toc197161403 _Toc197161445 _Toc197161404 _Toc197161446 _1.1_What_is _1.1.1_Deep_v _Toc197161447 _Toc185067957 _Toc197161448 _Toc197160621 _Toc197161408 _Toc197161450 _Toc197160622 _Toc197161409 _Toc197161451 _Toc197160624 _Toc197161411 _Toc197161453 _Toc197160626 _Toc197161413 _Toc197161455 _Toc197160636 _Toc197161423 _Toc197161465 _Toc197160637 _Toc197161424 _Toc197161466 _Toc197160639 _Toc197161426 _Toc197161468 _Toc197161469 _Toc129689420 _Toc197161470 _Toc188354585 _Toc188354712 _Toc188354966 _Toc188355087 _Toc188354587 _Toc188354714 _Toc188354968 _Toc188355089 _Toc188354589 _Toc188354716 _Toc188354970 _Toc188355091 _Toc188354591 _Toc188354718 _Toc188354972 _Toc188355093 _Toc188354593 _Toc188354720 _Toc188354974 _Toc188355095 _Toc188354595 _Toc188354722 _Toc188354976 _Toc188355097 _Toc188354597 _Toc188354724 _Toc188354978 _Toc188355099 _Toc188354599 _Toc188354726 _Toc188354980 _Toc188355101 _Toc188354601 _Toc188354728 _Toc188354982 _Toc188355103 _Toc188354603 _Toc188354730 _Toc188354984 _Toc188355105 _Toc188354605 _Toc188354732 _Toc188354986 _Toc188355107 _Toc188354607 _Toc188354734 _Toc188354988 _Toc188355109 _Toc188354609 _Toc188354736 _Toc188354990 _Toc188355111 _Toc188354611 _Toc188354738 _Toc188354992 _Toc188355113 _Toc188354612 _Toc188354739 _Toc188354993 _Toc188355114 _Toc188354613 _Toc188354740 _Toc188354994 _Toc188355115 _Toc188354616 _Toc188354743 _Toc188354997 _Toc188355118 _Toc188354618 _Toc188354745 _Toc188354999 _Toc188355120 _Toc188354623 _Toc188354750 _Toc188355004 _Toc188355125 _Toc188354624 _Toc188354751 _Toc188355005 _Toc188355126 _Toc188354625 _Toc188354752 _Toc188355006 _Toc188355127 _Toc188354627 _Toc188354754 _Toc188355008 _Toc188355129 _Toc188354628 _Toc188354755 _Toc188355009 _Toc188355130 _Toc188354653 _Toc188354780 _Toc188355034 _Toc188355155 _Toc188354660 _Toc188354787 _Toc188355041 _Toc188355162 _Toc188354662 _Toc188354789 _Toc188355043 _Toc188355164 _Toc188354664 _Toc188354791 _Toc188355045 _Toc188355166 _Toc188354666 _Toc188354793 _Toc188355047 _Toc188355168 _Toc188354668 _Toc188354795 _Toc188355049 _Toc188355170 _Toc188354670 _Toc188354797 _Toc188355051 _Toc188355172 _Toc188354672 _Toc188354799 _Toc188355053 _Toc188355174 _Toc188354674 _Toc188354801 _Toc188355055 _Toc188355176 _Toc188354677 _Toc188354804 _Toc188355058 _Toc188355179 _Toc188354679 _Toc188354806 _Toc188355060 _Toc188355181 _Toc188354681 _Toc188354808 _Toc188355062 _Toc188355183 _Toc188354683 _Toc188354810 _Toc188355064 _Toc188355185 _Toc188354685 _Toc188354812 _Toc188355066 _Toc188355187 _Toc188354687 _Toc188354814 _Toc188355068 _Toc188355189 _Toc188354689 _Toc188354816 _Toc188355070 _Toc188355191 _Toc188354691 _Toc188354818 _Toc188355072 _Toc188355193 _Toc188354693 _Toc188354820 _Toc188355074 _Toc188355195 _Toc188354695 _Toc188354822 _Toc188355076 _Toc188355197 _Toc188354697 _Toc188354824 _Toc188355078 _Toc188355199 _Toc109929154 _Toc112521811 _Toc197161471>>>OOO888n3;ÍcccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccJ  !"#$%&'()*+-,./0123456789:;<=>?<<<NNN[[[AAA<0I=ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccmmmJ4C&"DEl~"FGlHLIJTKd"LM "NOe"PlE"QR3 ST_U!VDy"Wk"XY }"ZeH[T!\h"]^\$_`Q"ab\cP!dh!efg4&"hj!iK"j̎!ksltm$unuowplxqyrys4ztzud{v|..5QQSSUl2%EE33\\   77J      !"#$%&'()*+,-./01234??QQSSUs;+JJ55_a   <<J  !"#$%&'()*+,-./01239*urn:schemas-microsoft-com:office:smarttagsState:*urn:schemas-microsoft-com:office:smarttagsStreet;*urn:schemas-microsoft-com:office:smarttagsaddress=0*urn:schemas-microsoft-com:office:smarttags PlaceType=1*urn:schemas-microsoft-com:office:smarttags PlaceNameB3*urn:schemas-microsoft-com:office:smarttagscountry-region8&*urn:schemas-microsoft-com:office:smarttagsCity94*urn:schemas-microsoft-com:office:smarttagsplace C_43410433334334&443443433344434343414034&4&43344FNWc''0'='126668>8*9.9::FF8IAIaaaazz{{O}W}_}j}s}v}*6ˊԊLSTWX]^dhp{|ɱձ=A      k o             18Rc. %ch Zf# u{  ""##%&()+,9DGJ^knS**//89<<j?p?![&[ccppO}j}k}r}-36>tϝѝ{|39BGY{ +2:#(  G T       u    <B_%>Olo  ""##%&()+,9DGJ3333333333333333333333333333333333333333333333333>81/0112722223334466;<yyz{||||O}k}~[i/;?msÍt{|?oF([$7+/cm    ""##%&()+,47GJ  ""##%&()+,GJ  A!gH _7R|dy 47BL/R4Tn4a5.w/OK P L{ P  ;g`N@tlujxl{ DxN.qpr}w ^`OJQJo( hh^h`OJQJo( SS^S`o(hH. pp^p`hH. @ L@ ^@ `LhH. ^`hH. ^`hH. L^`LhH. ^`hH. PP^P`hH.  L ^ `LhH.h ^`OJQJo(h rr^r`OJQJo(oh B B ^B `OJQJo(h   ^ `OJQJo(h ^`OJQJo(oh ^`OJQJo(h ^`OJQJo(h RR^R`OJQJo(oh ""^"`OJQJo(h ^`hH.h ^`hH.h pLp^p`LhH.h @ @ ^@ `hH.h ^`hH.h L^`LhH.h ^`hH.h ^`hH.h PLP^P`LhH.h^`OJQJo(hHh^`OJQJ^Jo(hHohpp^p`OJQJo(hHh@ @ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohPP^P`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohpp^p`OJQJo(hHh@ @ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohPP^P`OJQJo(hHh 88^8`hH.h ^`hH.h  L ^ `LhH.h   ^ `hH.h xx^x`hH.h HLH^H`LhH.h ^`hH.h ^`hH.h L^`LhH.h^`OJQJo(hHh^`OJQJ^Jo(hHohpp^p`OJQJo(hHh@ @ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohPP^P`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohpp^p`OJQJo(hHh@ @ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohPP^P`OJQJo(hH^`OJQJo(hH88^8`OJQJ^Jo(hHo^`OJQJo(hH  ^ `OJQJo(hH  ^ `OJQJ^Jo(hHoxx^x`OJQJo(hHHH^H`OJQJo(hH^`OJQJ^Jo(hHo^`OJQJo(hH hh^h`o(hH. @@^@`o(hH.0^`0\]^JaJo(hH.. ``^``o(hH... ^`o(hH .... ^`o(hH ..... ^`o(hH ......  `^``o(hH.......  00^0`o(hH........h ^`OJQJo(h rr^r`OJQJo(oh B B ^B `OJQJo(h   ^ `OJQJo(h ^`OJQJo(oh ^`OJQJo(h ^`OJQJo(h RR^R`OJQJo(oh ""^"`OJQJo(hh^h`OJQJo(hH^`OJQJ^Jo(hHopp^p`OJQJo(hH@ @ ^@ `OJQJo(hH^`OJQJ^Jo(hHo^`OJQJo(hH^`OJQJo(hH^`OJQJ^Jo(hHoPP^P`OJQJo(hHh^`OJQJo(hH^`OJQJ^Jo(hHopp^p`OJQJo(hH@ @ ^@ `OJQJo(hH^`OJQJ^Jo(hHo^`OJQJo(hH^`OJQJo(hH^`OJQJ^Jo(hHoPP^P`OJQJo(hH^`o(.hhhhhhhh^`OJQJo(hH88^8`OJQJ^Jo(hHo^`OJQJo(hH  ^ `OJQJo(hH  ^ `OJQJ^Jo(hHoxx^x`OJQJo(hHHH^H`OJQJo(hH^`OJQJ^Jo(hHo^`OJQJo(hH Dx{ P Px Ay ;g`gH Tn47pr}a5/O7BL/tlu                                    h=n! o?(da3+CNIV///000070F0j0k0|000011101E1J1e1f1q1u1y11111111112#23272u2v22222 3 333303F3G3d3333333333359x/y0yYyyyyyyyyyzyzzz{9{{||||}N}O}k}}c~~~~~~~~~~uvUVWXYZ[iՁ /;Ăłmۃ2o-!A[\EFSXl{KCv'(6EMQR[u"#$7b+-/`a   "%(+7GJ U/ykYA U};@@IP@UnknownholmesstaffGz Times New Roman5Symbol3& z Arial;SimSun[SO5& zaTahoma?5 z Courier New;Wingdings"1hFFv&)8)8!4dGG 2qHX ?CN2 Introduction holmesholmesP            Oh+'0  < H T `lt|Introduction holmes Normal.dotholmes2Microsoft Office Word@^в@.@,O<@,O<)8՜.+,D՜.+,L hp  university of sussexG Introduction Title` 8@ _PID_HLINKSA<P0)http://ideas.repec.org/s/wbk/wbrwps.html -.http://ideas.repec.org/p/wbk/wbrwps/3458.htmlF<*9http://www.unido.org/file-storage/download?file_id=60028-|' http://ssrn.com/abstract=6049924 _Toc1971614714_Toc1971614704_Toc1971614694_Toc1971614484_Toc1971614474_Toc197161446  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~Root Entry F0<Data .1Tablex'WordDocument6SummaryInformation(DocumentSummaryInformation8CompObjq  FMicrosoft Office Word Document MSWordDocWord.Document.89q