Countries, firms and global value chains.

Transcription

1 Countries, firms and global value chains. The internal and external coherence of the Trans-Atlantic Trade and Investment Pact. Abstract Giorgia Giovannetti 1 and Maria Luigia Segnana 2 (June,2015) The global economy is increasingly structured around global value chains (GVCs) that account for a rising share of global production, international trade, investment and global GDP. Studies from a range of disciplines show that global value chains, i.e. mechanisms through which the process of specialization is organized with goods being processed, and value being added, in multiple countries, are becoming the way production is organized at world level and many international reports make use of them in tracing the shifting patterns of global production. The emergence of GVC has changed the context and highlighted that gross trade data can be misleading on how value added is exchanged between countries. Recently many international organizations have produced and utilized the OECD/WTO TiVA database, to shed light on the scale, nature and consequences of this production networks. We have learned that gross trade can significantly differ from trade in value added, a fact earlier based on anedoctical evidence, but now quantified (and quantifiable). A gap that marks for differences in the global supply chain activity, across countries and over time. At the same time, supply chain analysis is the basis of trade agreements in progress, like the Trans-Atlantic Trade and Investment Partnership (TTIP) in a negotiation phase between the European Union and the United States. It is assumed that this agreement, similarly to others in progress, should have the largest impact if it covers as many dimensions of GVCs as possible. With the lenses of the economic and the business perspectives, this paper explores the coherence of production networks with trade memberships, by looking at external and internal coherence of the TTIP. First, after a short comparison of the positive but small estimates of the TTIP-gains from trade, the paper looks at economy-wide impact in scenarios of deeper integration. Second, the internal coherence of the on-going TTIP is examined, by looking at whether or not it is inclusive of small-medium enterprises. Third, the external coherence between the multilateral trading system and TTIP is scrutinized. Results show first that the internal and external coherence of TTIP are not so clear-cut; that the Partnership (as well as other current negotiation, including the EU-Korea and US-Korea) are not designed in a way to account for the GVCs and that, in any case, newly harmonized global supply chain discipline is not necessarily inclusive of small-medium enterprises. The case of SME in Southern Europe is an interesting example to check whether or not TTIP induce or improve SME s participation to global value chains (or what should be changed for it to do so). Keywords: Regional economic integration, free trade areas, global supply chain. JEL codes: F12, F13, F14, F15, O2. 1 Giorgia Giovannetti, Dept. of Economics and Management, University of Firenze, Firenze and European University Institute, Robert Schuman Centre, San Domenico di Fiesole, Italy. 2 Maria Luigia Segnana, Dept. of Economics and Management, University of Trento, Trento, Italy, marialuigia.segnana@unitn.it 1

2 1. Introduction 2. The methodological and economic assessments of TTIP. 2.1 The assessment 3. The internal and external coherence 3.1 The internal coherence 3.2 The external coherence : TTIP and the Mediterranean countries 3.3 The external coherence: Regional and multilateral agreements. 4. Concluding remarks References 1. Introduction The international economy is increasingly structured around global value chains 3 (GVCs) that account for a rising share of global production, international trade, investment and global GDP. Studies from a range of disciplines show that global value chains, i.e. mechanisms through which the process of specialization is organized with goods being processed, and value being added, in multiple countries, are becoming the way production is organized at world level 4. Many international reports make use of them in tracing the shifting patterns of global production 5. The emergence of GVC has changed the mapping of industrial and trade landscape, changing our perception of international trade, focusing upon trade in value added with an important impact on trade statistics and policy assessments. Two approaches have been used to measure trade in value added: the top down and the bottom up 6. From the top down, international input-output tables combine now national accounts with trade statistics and decompose of gross trade into its foreign and domestic value added contents 7. The results show a mapping of world changing trade landscape based upon country and sectoral value added data. Estimates 8 of trade measured in value-added terms show that almost 30 per cent of total trade consists of re-exports of intermediate inputs, thus indicating increased 3 We refer here to as global supply chains, global value chains, international production networks, vertical specialization, offshore outsourcing and production fragmentation. 4 Gereffi and Lee(2012). 5 OECD, 2013; OECD 2014, WTO and IDE- JETRO, 2011 amongst others 6 We do not consider here the approach based upon trade statistics on intermediates in foreign trade 7 Johnson(various years), Johnson and Noguera(various years), Koopman et al.(2010),foster et al.(2013),maurer and Degain(2010) 8 WTO(2013). 2

3 international interdependence through international production chains. Since the mid-1990s, this measure has risen by almost 10 percentage points. These new statistics can surely help designing better trade policies with new angles for trade analysis (country s place in GVC, bilateral trade and revealed comparative advantage), new answers about how supply chains have changed patterns of international trade and new needs for adequate links between regulatory trade regimes and business reality. But global supply chain activities, were also examined from the bottom up, with many case studies 9 showing the geographical decomposition of a product value into the components and services used for its production. Direct measurement or collection of data on supply chains at customs frontier or through supply chain software show micro data for a single product or a single sector in industries such as electronics, apparel, and motor vehicles and trace products from origin to final destination. Accessing these data could be difficult and they are often based on hoc surveys. We are in presence of two mappings of the world changing trade landscape: one based upon aggregate value added data that question traditional trade data and propose a substantial revision 10 and one base upon partial but ad-hoc micro analysis of global supply chain. Integration between the top down and the bottom up approach is difficult as well as integration between GVCs ad hoc surveys and Melitz s empirics. In recent reports of international institutions 11, global value chain is offered as the tool to explain and justify oncoming trade agreements, that will rationalize trade between different economic areas, as they will deepen, enlarge and consolidate value chains. Non tariff-measures (NTMs) and the actionability of their reduction/harmonization do play the leading role for the harmonization of supply chain disciplines. This is often invoked even in the case of Transatlantic Trade and Investment Partnership (TTIP), now in a negotiation phase between the European Union and the United States. It is assumed that this agreement, similarly to others in progress, would have a larger impact if it would cover as many dimensions of GVCs as possible and if it would be signed thinking Global Value Chain 12. With the lenses of the economic and business perspectives, this paper explores the coherence of production networks with trade memberships, by separately considering external and internal coherence of a specific trade agreement: the TTIP. Section 2, after a short comparison of the positive but small estimates of the TTIP-impact, describes economy-wide impact in scenarios of this deep 9 See the many examples from Center on Globalisation, Governance and Competitiveness at Duke University 10 Note that differences can be substantial. For instance, WTO 2013 shows that the bilateral current account deficit of US versus China is not at all large if data are considered in value added. 11 Examples are EU(2012), UNCTAD(2013), WTO (2011). 12 Hoekman,

4 integration. Section 3 examines the internal and external coherence of the ongoing TTIP, explaining what we mean by that. More specifically, internally, or for the EU s industrial structure in 3.1, by looking at the impact on SMEs in order to check whether or not TTIP is inclusive of small-medium enterprises. Externally, in 3.2 and 3.3 by looking at the trade diversion and at the link between the objectives of the multilateral trading system and those of the TTIP. Results show that the internal and external coherence of TTIP is not so clearcut, especially because of GVCs. On the one side, GVCs enhancements justify the increase of the economic interests at stake in the TTIP. On the other side, the present benefits of the agreements are measured on gains that do not have value chains as their starting point, as shown in section 2. Furthermore, the harmonized global supply chain discipline that TTIP should aim at, through reduction of regulatory barriers is not necessarily inclusive of SMEs, which could represent a serious issue especially for Southern European countries, which are characterized by a larger share of SME. 2. The methodological and economic assessments of TTIP The economies of the European Union and the United States are very important trading partners for each other. Together they account for 40% of total global trade (more than $1.5 billion in transatlantic trade every day, source EC, 2015). The $3.75 trillion EU-U.S. transatlantic economy employs 14 million workers on both sides of the Atlantic. Since 2001, Europe has accounted for roughly twothirds of total global investment flows into the U.S. by far the most significant source of foreign investment in the U.S. economy. The EU is also the largest source of US Imports and second largest US export market. European companies are the leading foreign investors in the U.S. and American companies invest far more in EU countries than in Asia. Although average tariff levels are already relatively low 13, various non-tariff barriers or NTBs (often in the form of domestic regulations) on both sides of the Atlantic constitute important impediments to deepening transatlantic trade and investment linkages. The first proposals for a Transatlantic Free Trade Area date back to the 1990s. But only in February 2013, also as a reaction to the stalemate of multilateral negotiations within the WTO, and after important preliminary work throughout 2012, the launch of the Transatlantic Trade and Investment Partnership (TTIP) was announced. The TTIP is a trade agreement presently being negotiated between the European Union and the United States, with a final agreement that would grouped together 3 parts: Market access, (better access to the US market), Regulatory cooperation (cutting red tape and costs) and Rules (new rules to make it easier and fairer to export, import and invest). The first round of negotiations 13 Estimates are that tariffs are on average around 2.2% in the US and, 3% in EU, even though with substantial differences amongst different sector. 4

5 was in July 2013; as of April 2015, nine rounds have been completed but the completion date, previously announced for the end of 2014, is still unknown 14. The main aims of the partnership are to increase trade and investment between the US and EU by reducing tariffs (particularly on agricultural products), aligning regulations and standards, improving protection for overseas. Besides cutting tariffs across all sectors, the EU and the US aim at tackling barriers behind the customs border such as differences in technical regulations, standards and approval procedures. Averaging around 3%, tariffs between the EU and US are already low, and both sides foresee their possible elimination under the Agreement. Hence,the negotiating energy is devoted to reducing non-tariff barriers to trade, with the aim of harmonising product regulation and standards (e.g. labelling, product specifications, sanitary requirements) in areas where these are deemed necessary, and eliminating them in areas where they are unnecessary. Other areas being contemplated include protection for foreign investors, co-operation to achieve greater participation by SMEs in EU-US trade, provisions on intellectual property. Despite the initial scarcity of information, controversial points were those on food standards, public procurement, intellectual property, air and maritime transport, financial services 15,and today, on investor-state dispute settlement (ISDS ). At first, information was not so easily available, but after the consultation on investment protection 16 and the online survey on Small and Medium Sized Enterprises 17, it has improved together with the diffusion of EU s position. On regulatory cooperation, the EU has an updated proposal 18, but the balance between the protection of investors and the protection of the States ability to regulate in the public interest(the ISDS 19 ), is still under discussion 20. The potential of TTIP in terms of mutual economic benefit represents a long term strategy with which EU and US could recover from the financial crises in a situation of increasing international competition with emerging economies. They could also recover a leading role at the international level. Furthermore, TTIP can be seen as the response of Western economies to potential trade diversion effects, given the important regional agreements in progress in Asia. On this basis, an overview of existing assessments and findings of TTIP s effects is useful to check whether or not they can offer a suitable basis for this trade reform. 14 See for description of the different rounds of negotiations as well as for updates 15 Lannoo (2013) may/tradoc_ pdf 20 Bruhn (2015) develops an econometric test about the importance of global value chain trade and regulatory differences in explaining the likelihood of a country pair to include an (enforceable) investment provision in the PTA. 5

6 2.1 The assessment Given the size of the economies involved, and the ambition of the agreement, the welfare gains for US and EU are predicted to be small but positive, even if they vary depending on the nature of the agreement that might be reached. TTIP will also have important consequences for outsiders, not necessarily positive. As far as EU and US are concerned, the welfare gains given by the additional trade created are likely to overcome the welfare losses due to trade diversion, but differ substantially between and within EU countries and US states depending on their sectoral and skill composition. A number of studies find that there are mutual economic benefits from trade liberalization. In the debate, a few selected studies, Ecorys (2009), CEPR (2013), CEPII (2013) and Bertelsmann/IFO (2013) have set the tone, suggesting that effects are positive on both sides of the Atlantic 21. All studies simulate various scenarios by comparing policy changes to a baseline calibration, with a forecast period of on average 10 years. The aim is to identify and analyse existing NTMs at the sectoral level. They then consider the potential economic impacts the alignment of these measures could have at the sectoral level as well as overall for US and EU. Any substantial impact of such an agreement on trade flows for the signatories and for third markets is driven by changes in NTMs. However, analyzing the effect of NTMs on trade is difficult due to the breadth of policies covered as well as to the difficulties in measuring. All studies report small but positive effects on GDP trade flows and real wages in the EU. GDP and real wages increases are estimated by most studies to range from 0.3 to 1.3 per cent, even in the most optimistic liberalization scenarios. Three of these studies 22 provides computable general equilibrium (CGE)-based estimates for the economy-wide impact of reducing both tariff and non-tariff barriers (NTBs). Estimates are provided with regards to expected changes in GDP 23, sector output, aggregate and bilateral trade flows, wages, and labour displacement, among other issues. The analyses use the GTAP7 and 8 database (projected to 2027), in conjunction with NTB estimates reported in the Ecorys (2009) study. Different policy options are examined for the deepening of the bilateral trade and investment relationship between the EU and US. Options go from partial agreements that are limited in the scope of barriers they would address (tariffs only, or services 21 These studies are reviewed in OFSE(2014). Using the United Nations Global Policy Model, Capalbo (2014) simulates the impact of TTIP on the global economy in a context of protracted austerity and low growth especially in the EU and US. His results differ significantly from the results of the four studies reported here and highlight a likely negative impact on GDP and employment of several countries, including the US and EU. 22 We do not discuss the specific methodological choices of the four studies. 23 For instance, the most widely quoted, CEPR (2013), finds that an ambitious and comprehensive trade and investment agreement could bring aggregate economic gains of 119bn per year to the EU (0.9% GDP) and 95bn (0.8% GDP) to the US. Bertelsmann Foundation estimated that a comprehensive deal would raise per capita incomes in the EU by 5% and the US by 13%. 6

7 only, or procurement only) to a full-fledged free trade agreement. Given the similar data base and the closely related methodological approaches, these three studies report gains in real income and trade flows within a similar range for all participating countries. In contrast Bertelsmann/IFO 24 finds the most pronounced benefits and correspondingly penalization for outsiders, because of trade diversion. Scenarios go from a less ambitious agreement that includes a 10 per cent reduction in trade costs from NTBs and nearly full tariff removal (98 per cent of tariffs) to a more ambitious scenario that includes the elimination of 25 per cent of NTB related costs and 100 per cent of tariffs. In both scenarios more ambition is imposed on the lowering of procurement-related NTBs than for other NTBs affecting goods and services. It is assumed that NTBs linked to procurement are reduced by 25 per cent or 50 per cent, in the less ambitious and in the ambitious scenarios respectively. The results 25 indicate positive and significant gains for both economies. Under a comprehensive agreement, GDP is estimated to increase by between 68.2 and billion euros for the EU and between 49.5 and 94.9 billion euros for the US (under the less ambitious and more ambitious scenarios). However, if the agreement would be confined to tariff liberalisation only, or services or procurement liberalisation only, the estimated gains would be significantly lower. The overall message is that negotiating an agreement that would be of a comprehensive nature would bring significantly greater benefits to both economies. However, attention is necessary whenever NTBs are considered, because they are difficult to measure and their reduction as well the channel of transmission of their effects can be questioned. A first caution is in the use of general equilibrium models to simulate the impact of economic policies. Within these models, the GTAP 26 is certainly the workhorse of all simulations of the effects of trade policies. But non-tariff barriers (NTBs) require some extra attention in the case of GTAP models. These models are traditionally suitable to study the direct impact on the prices of trade policies rather than the indirect impact through the abolition of non-tariff barriers. There are many good reasons for the lack of global analyzes or simulations on non-tariff barriers 27. Under a common name, NTBs mean a plethora of commercial tools, potentially distorting. Unlike tariffs, NTBs are difficult to quantify and the relevant information difficult to collect. UNCTAD, for example, 24 Based upon a gravity model. 25 A comparison is reported in table 1, from the OFSE s comparative analysis. 26 The GTAP database (The Global Trade Analysis Project is the result of a research network that conducts quantitative analyzes on issues of international economic policy, also used by the World Bank. It is coordinated by the Center for Global Trade Analysis at Purdue University and at the base there is a computable general equilibrium model. One of the main areas of application involves simulation of the effects of trade agreements. 27 See, for an overall assessment Fugazza and Maur (2008). They point out that, among the lessons that are derived from various simulations of different treatments of NTBs in CGE models efficiency gains caused by the elimination of NTBs are likely to be overestimated. 7

8 through the Trade Control Measures Coding System, identifies more than 100 instruments grouped into six categories. A second concern regards how NTBs are defined and estimated. Ecorys estimates of NTB, comprehensive of regulatory divergence, are built on firms perceptions about the restrictiveness of these measures. They are a multiple of 3 per cent of Anderson and van Wincoop (2004) and obviously, the higher the NTBs to be removed the higher the potential gain from trade. A third concern is the common hypothesis about the effects of NTB reductions via prices changes and the assumptions of price elasticities. Price elasticities, which determine the quantitative reaction of demand and supply are high in these studies and double the size compared to the macroeconomic literature. It is rather obvious that higher assumed value for elasticities drive the gains from trade. Even abstracting from NTBs, further observations about whether or not existing assessments of TTIP offer a suitable basis for such an important trade reforms can be shortly summarized. For instance, comparing other assessments of trade liberalization experience (e.g. NAFTA), we find that even the NAFTA studies were based on the ex-ante kind of CGE modeling used for TTIP. A significant gap exists between ex-ante projections and ex-post evaluations with regard to NAFTA s effects on GDP, wages and employment. Ex-post analysis of the NAFTA impacts suggest that exante impact projections substantially overestimated the economic effects 28. But attention must be paid even to adjustment costs that can be substantial and are mostly neglected in the TTIP studies: social costs of regulatory changes and harmonization of laws, procedures and standards often safeguard collective preferences 29. All studies, but particularly the Ecorys study, assume that a reduction of NTMs is welfare-enhancing, ignoring that non-tariff measures pursue public policy goals. Trade and welfare effects of NTB do not necessarily carry the same sign because the effects depend on the nature of the market failure that the measure addresses. The elimination or alignment of an NTM thus could imply a social cost for society with both a short-term adjustment cost for public institutions and for firms required to align their administrative procedures, production processes and products to the new standards, for consumers with new information costs for society insofar as this elimination threatens public policy goals (e.g. consumer safety, public health, environmental safety), which are not taken care of by some other measure or policy. Furthermore, it is assumed that around 50 % or 25% of all existing NTMs between the EU und the US are actionable, while CEPR assumes a 25% actionability level. This includes sensitive sectors such as foods and beverages, chemicals, pharmaceuticals and cosmetics or automotives, where the safeguarding of public policy goals is perhaps most crucial. Could these high levels of actionability be implemented 28 For a review and a comparison of this extensive literature see OFSE (2014) 29 Macroeconomic adjustments costs are also analyze in OFSE,

9 without any losses to the quality of regulation in the public interest? Surely, the incurred social costs of TTIP regulatory change might be substantial, and require careful case-by-case analysis. In sum, the review of the assessments suggests that significant gap exists between ex-ante projections and ex-post evaluations with regard to NAFTA s effects on GDP, wages and employment. Some of them are inherently related to the difficult estimates of NTB and their actionability. Some others focus upon the neglected costs of the invoked harmonization. Some concentrate on the tendency to overestimate the benefits and underestimate the costs of freertrade. The overview of existing assessments and findings of TTIP show that its small but positive effects are derived without any references to GVCs. They can be treated with scepticism: for the many specific reasons that have mentioned, especially those related to simulating an uncertain future on the basis of questionable assumptions. But they can be treated with scepticism also because they are derived without any references to GVC analysis, that was instead used as the starting and motivating point. The value chains have taken now a strange role: on one side they are used to justify the performance of the economic interests at stake in the TTIP. On the other hand, the actual benefits of the agreements are measured on gains that do not have value chains as their starting point. 9

10 Table 1: Overview on basic assumptions and findings Source: OFSE(2014). 10

11 3. The internal and external coherence of TTIP Given the features of the agreement and the macro assessment of its impact, the question of to what extent the agreement can coordinate, strengthen and make consistent trade between the two geographic areas with its underlying chains value is still alive. Such a question about its internal coherence cannot be disentangled by the complementary question about its external coherence, or the coherence with respect to the existing multilateral trading system. In search for an answer, it is necessary to recall the changing features of trade agreements in the 21th century. The coordination problem that a trade agreement should address is well known. But old agreements have exchanged "market accesses" where the exchange was "my market in exchange for your market". The new or more recent agreements first enriched and then definitely extended the coverage of these coordination mechanisms. In fact, in the presence of international value chains and thus the possible relocation of production, the exchange is no longer an exchange of liberalization commitments or of "market accesses": it is an exchange that provides and requires protection and liberalization together. Or, virtually delocalized systems and technologies (tangible and intangible assets) adequately protected and the results of activities (e.g the exchange in final goods) adequately liberalized (Baldwin, 2011, 2012). We are in presence of new features of exchanges(baldwin and Lopez-Gonzales, 2013), governed by different coordination mechanisms, which require new approaches and new policy areas to consider (Hoeckman, 2012). Such a coordination problem could find a solution as much commercial interests (the size of the value chain) and extension of the preferential agreements overlap. In this case, we could expect a solution to the co-ordination problem and, therefore, a success of the trade agreement, with the standard static and dynamic gains from trade. 2.1 The internal coherence The CEPR study (2013) points to the importance of the elimination of NTB ontariff barriers that would count for about 80% of gains and would result in a reduction between 10 and 20 per cent in the prices of products, and (limited) job creation. The internal coherence of TTIP can be evaluated within the EU, by looking at the trade and industrial impact of non-tariff measures (NTM) reduction. First, one of the most obvious impact is the change in the EU_US trade flows. The overall positive impact of TTIP on total exports is very concentrated and conceals large trade diversion 30. For example, the total would be very concentrated sectorally as the increase of the EU-US bilateral trade would be substantial, especially in the field of motor vehicles and their engines, reflecting the importance of trade in parts and components and the significant technical barriers or non-tariff that characterizes it. Among other things, the aggregate 30 Results on trade diversion are different for different scenarios in the studies on TTIP 11

12 earnings are also based on the potential increase in exports outside the European single market: this would include, for example, the automotive sector (41%), metal products (12%), products of the chemical (9%). In sum, data seems to suggest a possible strengthening of the German production platform. Taken for given the small but positive aggregate results and the change in the trade flows, a second impact of the non-trade measures elimination could be very unevenly distributed in presence of the heterogeneous landscape of European industrial system across countries, sectors, firms size, firms characteristics and transatlantic interests. For example, at the country level, firms in Italy and Spain, are on average 40 percent smaller than firms in Germany. The low average firm size translates into a chronic lack of large firms. In Italy and Spain, a mere 5 percent of manufacturing firms have more than 250 employees, compared to a much higher 11 percent in Germany. And large firms contribute disproportionately to the economic performance of countries: usually, they are more productive, pay higher wages, enjoy higher profits and are more successful in international markets. In order to explain the successful story in international markets, the literature has drawn attention to the fact that exporting and non-exporting firms coexisted in the same industry but were marked by clear defining characteristics. It is a combination of sunk costs and heterogeneity in the underlying characteristics of firms that explains why not all firms export. In the context of recent trade models with heterogeneous firms, conforming with regulatory standards in an importing country constitutes a fixed entry cost to penetrate that market, as discussed by Bernard et al. (2011), but may also constitute a variable cost that needs to be incurred every time the firm exports to that market, e.g., if more costly inputs need to be used to meet regulations. The effect of non-trade measures (NTM) on firms participation to international trade is not straightforward. NTM may reduce additional costs on exporters and here it is important to distinguish their effects 31 on fixed costs of producing from the effect of fixed costs of exporting and from the effect on variable trade costs. Considering also the demand side, it is hard to say a priori whether NTM have a positive or negative effects on exports and its intensive and extensive margins. For example, focusing on a model of a single firm making export decisions, Chen et al. (2008) allow compliance with standards to impose additional production costs on firms but also to possibly have a positive effect on demand (in terms of consumers willingness to pay for the products). The net effect of standards on a firm s choice of optimal scale and export scope depends on the strength of the standards -induced increase in costs versus the strength of the standardsinduced increase in demand. Looking at the effects of NTMs on extensive and intensive margins of export, the effects of SPS and TBT can be very diverse. SPS measures largely influence the variable trade costs, and TBTs mostly increase fixed trade costs with different 31 Fixed trade costs affect the extensive margin, i.e., a firm s decision of whether to export to a market, as they do in the seminal model of Melitz. Fixed trade costs, however, do not affect the intensive margin, i.e., the firm s exports to that market, in those models whereas variable trade costs affect both margins of firm exports. 12

13 consequences on the trade margins. Even the level of aggregation/disaggregation matters because there is substantial heterogeneity of response of exports to NTBs across product lines for more disaggregated data. In presence of differential effects on margins of trade of different non trade measures, what effect can we expect from NTM elimination on extensive and intensive margins of trade in countries characterized by very different distribution of firms size? Let us simply list some examples on SMEs showing (1) the importance of SMEs in EU exports to the US and their reported trade barriers; (2) the differential importance and impact of technical standards by firm s size and (3) the positive impact of participation in global value chain on the propensity to internationalize. These three cases have something in common: they focus on SMEs and their perceived barriers, on the differential impact of NTMs harmonization, on the impact of GVC participation on the propensity to internationalize. In the first two cases the results show that regulatory harmonization is increasingly important in firm size and that the effects might be different by firm s size. The last case is useful in order to remind us about the importance of the GVC participation. The first case is taken from the recent Report on Small and medium Sized Enterprises and the TTIP, a report with a clear focus on how TTIP makes trade easier for SMEs. By looking at SME s role in EU export to the world and to the US, the Report shows the results of a survey on the most significant NTB in the US. The setting looks at the USA-EU s trade by SMEs, often underestimated even though 28% 32 of the EU direct exports to the US are by SMEs and for which countries like Italy Spain have a proportion and export value of exporting SMEs to the US which is above the European average 33. The results of the survey show that the level of exporting activity increases with firm size, that in case of goods the most significant NTBs are SPSs and TBTs and that they are increasingly important in firm size; that in case of services, the existing legal limits on the movement of people are particularly detrimental to smaller business interests, as they cannot rely on locally hired personnel. The second case shows that regulatory harmonization or product standardization is quite differently perceived by firms of different size. The harmonization is often considered a form of standardization similar to the technology transfer that affects total factor productivity. But it is perceived differently by firms of different sizes, and some studies show that larger firms perceive it as a net benefit, while smaller companies find it hard to perceive it as a benefit altogether 34. International product standards can be difficult to use, countries preferences and interests vary and the capacity to influence international standard setting may differ among countries. In this case, it is not entirely clear how the harmonization of regulatory barriers, can be transformed into opportunities to 32 European Commission, European Commission, Blind(2004). 13

14 reduce costs of participation in international markets for small-medium enterprises. The third case shows the positive impact of participation in global value chain on propensity to internationalize. One empirical study (Giovannetti, Marvasi and Sanfilippo, 2015), based on a sample of Italian exporting firms, find a positive correlation between the participation of SMEs in global value chains and the propensity to become an exporter. Even small and less productive firms, if involved in production chains, can take advantage of reduced costs of entry and economies of scale that enhance their probability to become exporters. The empirical analysis over 25,000 Italian firms, largely SMEs, include direct information on the involvement in supply chains and show positive and significant impact of being part of a supply chain on the probability to export and on the intensive margin of trade. The number of foreign markets served (the extensive margin), on the other hand, does not seem to be affected. Even being in different positions along the chain, i.e. upstream or downstream, matters and in the Italian case, downstream producers tend to benefit more. In sum, TTIP s assessments and related surveys generate: (i) (ii) the strengthening of the manufacturing German platform in terms of sectoral composition of the expected increase in exports; differential effects on different firm s size, because the importance of regulatory standards both as SPSs and TBTs is differently perceived by firms of different size and is more important for larger firms. At the same time, we are in presence of a positive link between GVC participation and internationalization choices (at least in the Italian case). To what extent newly harmonized value chains, reinforced by transatlantic cooperation could integrate SMEs is still an open question because of the different impacts and importance of NTMs by firm s size. 3.2 The external coherence: Regional and multilateral agreements TTIP raise problems of internal consistency and uneven impact. But it also raises a problem of external consistency, or coherence between the objectives of the multilateral trading system and those of the preferential agreements 35. In other words, the coherence between the multilateral system and the proliferation of agreements with a size such as to replace the multilateral system. In fact, there are different possible options 36, the potential complementarity or substitutability between a system of trans-oceanic or mega-deals and the 35 Antràs and Staiger (2012) indicate that the rise in offshoring and its implications for international price determination is likely to undermine the effectiveness of the pillars of the GATT/WTO framework. 36 Nakatomi,

15 multilateral trading system, locked in Doha. Complementary, because TTIP is an example of the new trade-features treated in a regional rather than multilateral context. Substitutability, because the transatlantic bargain is often referred to as a super-doha: it would provide for new use of most-favored-nation clause applied in a non-discriminatory basis to all countries wishing to join, if they were in line with the standards agreed by the members. The transatlantic bargain traditionally is as an example of whether or not the nations prefer to treat the new features in a regional rather than a multilateral agreement. And yet, the transatlantic bargain is referred to as a super-doha because it would provide a new use of most-favored-nation clause applied in a non-discriminatory basis to all countries wishing to join, if they were in line with the standards agreed by the members. Looking at the scenarios of these mega-deals compared to the multilateral system, at least two extreme cases can be exemplified: a first integrated scenario, in which the juxtaposition of the old (WTO) and new rules (associated with new agreements under negotiation) leads to a rapid and effective harmonization and contributes to the redefinition of the multilateral trading system. And a second scenario, highly fragmented, where the existing rules and those derived from virtually new agreements do not integrate, or create regulatory environments very different inside and outside the regional trade agreements. With significant effects when one considers that value chains can change frequently, by geographical area, by the number and characteristics of participants, effects that have already been reported in the literature as spaghetti bowl or noodle bowl 37. Between these two extremes there are some proposals that deserve attention. For example, the proposal 38 to take into considerations more precise issues related to value chains and proceed via specific topics rather than by contrast between regional and multilateral systems. It would be possible in this case to reach a plurilateral agreement (International Supply Chain Agreement or ISCA): its aim should be to search of the common denominator that affects the various regional or sectoral value chains. 3.3 The external coherence: TTIP and the Mediterranean countries According to three studies, TTIP benefits will come at the cost of reducing bilateral trade between EU Member States. In a deep liberalization scenario, intra-eu trade could fall by around 30 %. The reason is that the EU countries exports will be substituted for by cheaper extra-eu imports. In addition, diversion effects in global trade from TTIP could be harmful for developing 37 A reminder about the term that was used for the first time by Jagdish Bhagwati and Anne Krueger in 1995 and then widely spread and used. In its original meaning it is used as an inevitable consequence of the proliferation of different free trade areas and their characteristics selective. It refers to how the production of parts and components can take advantage of different locations in different areas of free trade using price differentiation that characterizes them in an attempt to export the finished products to consumers at the lowest countries' low. 38 Nakatomi(2012). 15

16 countries one study expects negative real GDP change of 2.8% for Latin America and 2.1% for Sub-Saharan Africa The impact of TTIP on total exports conceals large trade diversion effects. In particular, intra-eu trade diversion is reported by CEPR(2013) and overall, this negative trade diversion effects would considerably exceed trade creation effects (Bertelsmann/IFO). Negative trade diversion effects for third countries due to bilateral trade agreements and mutual recognition agreements are reported in Bertelsmann/IFO but, in contrast, CEPR (2013) identifies a positive impact of TTIP for all other regions in the world due to the inclusion of spillover effects. Also CEPII sees negative consequences for exports of selected ROW countries in their reference scenario, but a positive impact if spillovers are included Thus, the assumptions of spillover effects enable CEPR and CEPII to avoid a conflict with the EU s commitment to Policy Coherence for Development. These (unconfirmed) economic gains for EU and the US are likely come at a price, especially for neighboring countries. Despite a boost for EU and US, that can favourably impact the world GDP, countries not participating in the TTIP, especially those that are traditional trade partners of the US and EU, could face a decrease in real income and employment. For instance, Canada and Mexico, members of NAFTA but outsiders in the TTIP are expected to suffer a substantial GDP decline according to all the existing estimates. More specifically, countries with an existing Free Trade Agreement with EU or US would be affected by an additional form of trade diversion, what Deardorff (2014) defines reversal of trade diversion. On top of the conventional trade diversion effect, imports that were diverted from the new partner (US) by the first FTA revert to the new partner with the second FTA (TTIP). This could further enhance the already high existing inequalities between developed west and emerging economies. Mediterranean countries are likely to be highly penalized by the signing of TTIP. The majority of those EU's neighbors are rather small countries and have signed in the past Free Trade Agreement with the EU 39. We can summarize the trade effects of TTIP on these countries. 40 Albania, Serbia, Montenegro, Bosnia & Herzegovina and Macedonia share a Free Trade Agreement (FTA) among themselves - CEFTA - and with the EU but not with the US. They present similar characteristics: EU is their main export destination market and import origin, while their trade with the US is insignificant (Deardorff, 2014). These countries are likely to suffer from two forms of trade diversion. In the first place, the conventional trade diversion effect since their exports to US must now compete with tariff-free exports of the EU; in the second place, the trade that was originally diverted from US by the first 39 These countries which belong to CEFTA (Central Europe FTA) and the Barcelona process (North Africa and Middle East) have existing Free Trade Agreement with EU; the only exceptions being Moldova, Kosovo and Libya. 40 Deardorff (2014), on which our analysis partially builds, does not include Libya, which we consider, given its geographical proximity and its importance as trade partners for some EU countries. 16

17 FTA agreement between EU and CEFTA members now may revert back to US with TTIP. All countries within the Barcelona Process - namely Algeria, Egypt, Lebanon, Syria, Tunisia, Israel, Jordan and Morocco - have existing FTAs with EU. Israel, Jordan and Morocco also have either FTA or EIA with the US. The countries involved in a FTA only with EU has the EU as the main partner and also trade significantly with the US; these countries are likely to be penalized by substantial the trade diversion effect. Egypt is an interesting example: the estimates of Felbermayr et al (2013) predicts that Egypt will lose 2.8% in real income per capita in a broad free trade agreement scenario. Countries which have free trade agreements both with EU and US (i.e. Israel, Jordan and Morocco) will be the ones which lose the most. To the extent that their significant exports to both were diverted from direct EU-US trade, this advantage will vanish with TTIP. Furthermore, Mediterranean countries may face a decrease also in foreign direct investments, since investors from Asia or sovereign wealth funds may increase FDI in the EU (and US) as a mean of avoiding tariffs and non tariff barriers. Turkey however is a special case, among the Mediterranean EU's neighbors. In 2012, more than 40 percent of its foreign trade was with the EU and the U.S.; two-thirds of Turkish capital was invested either in the EU or the U.S. and it is in a Custom Union (CU) with the EU, negotiated with the expectation that it would be a transitional arrangement while Turkey moved towards full EU membership, and that in the meantime it could help Turkey to strengthen its economy. A closer transatlantic trade and investment partnership may adversely affect Turkey, more than other countries. Its CU with the EU, implies that Turkey shares EU's external tariffs which will presumably include its zero tariffs on imports from US under TTIP and that Turkey will not benefit from reduced US tariffs on its exports, since it is not formally part of EU. In other words, U.S. products would enter the Turkish market freely without duties, while Turkey would continue to face duties and other limitations in the U.S. market. The current deficit that Turkey had in its trade with the U.S. (totaling to around $8.5 billion in 2012, see Kirisci, 2013) would probably increase. Furthermore, it is likely that further trade diversion would also occur as South Korean, (thanks to the South Korea-U.S. Free Trade Agreement, or KORUS FTA) and TransPacific Partnership (TPP) countries goods are likely to enter the U.S. market preferentially 41 ( see Deardorff, 2014).. This peculiar situation is likely to make Turkey worse-off after TTIP: According to Felbermayr et al. (2013) Turkey will lose 2.8% in real income per capita. On the other hand, a possible inclusion of Turkey within the TTIP could increase GDP up to 4.6% and will also favor the EU and the US in terms of higher GDP growth rates. 41 The top export from Turkey to the U.S. (vehicles, machinery, iron and steel products, and cement) overlap to a large extent with exports of the EU, South Korea and some Asia-Pacific countries. This is likely to amplify the adverse impact. 17

18 Libya, Moldova and Kosovo are the only Mediterranean countries which have not yet concluded a Free Trade Agreement with EU. 42 The trade between Moldova and the EU is substantial, but more significant if we consider also Eastern European Countries and Russia, while Libya has EU as the main trading partner (in ,7% of imports and 78,4% of exports are from/to EU). As complete outsiders of TTIP and having EU as the main trading partners, these countries are likely to suffer from trade diversion. In conclusion, Mediterranean countries, are likely to lose if TTIP is implemented, unless they can successfully integrate their firms into global value chains. 4. Concluding remarks The economic results of the transatlantic bargaining are not clear-cut, especially because if trade effects go through the value chains, then the potential gains arising from trade agreements cannot be assessed independently from the supply chains. TTIP is an example where value chain is playing a strange role: on one side it is used to justify the convergence of the economic interests at stake in the TTIP between the two sides of the Atlantic. On the other side, the actual benefits of the agreements are measured on gains that do not rely on value chains. From a methodological point of view, in the presence of GVCs, we have learned that gross trade differ from trade in value added, a fact known for some time, but now quantified. A gap that marks for differences in the global supply chain activity, across countries and over time. We have learned that the improvement of trade statistics in terms of value added can create new links with supply chain analyses but data improvements are necessary in order to link and enhance micro and macro analysis. From the existing assessments of TTIP impact at the country and industrial level we show that they can be a reference point for this important trade reforms but three things have to be taken into account: the coordination that a trade agreement develops is different: it is now a mixing up of protection (Investment, technologies) and liberalization (trade flows) and it is in this perspective that the TTIP has to be considered. the coherence of TTIP impact with the enhancement of GVC of European firms can be a problem especially for SMEs. TTIP could be a substitute of the multilateral trading system or a new center of attraction for countries in lines with its standards. Hence, this agreement impacts on European industrial systems. 42 Deardorff (2014 ) does not include Kosovo in the analysis because of lack of data. 18