Local Content Requirements: Promises and Pitfalls

Ing, Lili Yan (Ed.); Grossman, Gene M. (Ed.) Book Local Content Requirements: Promises and Pitfalls Routledge-ERIA Studies in Development Economics Provided in Cooperation with: Taylor & Francis Group Suggested Citation: Ing, Lili Yan (Ed.); Grossman, Gene M. (Ed.) (2024) : Local Content Requirements: Promises and Pitfalls, Routledge-ERIA Studies in Development Economics, ISBN 978-1-003-80691-2, Routledge, London, https://doi.org/10.4324/9781003415794 This Version is available at: https://hdl.handle.net/10419/290617 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ As anti-globalization and geopolitical tensions continue to rise, the use of local content requirements (LCRs) around the world has become more noticeable than ever before. The reasons for adopting LCRs range from ensuring domestic supply availability, job creation, and increasing value added to safeguarding national security. Ing and Grossman examine country-specific as well as firm-product level exercises to explain how LCRs reduce fair competition, resulting in lower trade and productivity, which ultimately lowers world economic output and overall human welfare. Countries around the world are investigated with specific attention to the US, China, Indonesia, and resource-intensive countries, including mining-intensive ones. The book also presents productand firm-level analyses, answering the question of why countries adopted LCRs and how LCRs affect the world economy. This book is a useful resource that will interest policymakers, researchers, and advanced undergraduates interested in international trade, industrial policy, political economy, labour economics, and development economics. Lili Yan Ing is a lead advisor (Southeast Asia Region) at the Economic Research Institute for ASEAN and East Asia (ERIA). She also serves as secretary general of the International Economic Association (IEA). Gene M. Grossman is the Jacob Viner Professor of International Economics in the Department of Economics and the School of Public and International Affairs at Princeton University. Local Content Requirements Production Networks in Southeast Asia Edited by Lili Yan Ing and Fukunari Kimura The Indonesian Economy Trade and Industrial Policies Edited by Lili Yan Ing, Gordon H. Hanson, and Sri Mulyani Indrawati World Trade Evolution Growth, Productivity and Employment Edited by Lili Yan Ing and Miaojie Yu East Asian Integration Goods, Services, and Investment Edited by Lili Yan Ing, Martin Richardson and Shujiro Urata Developing the Digital Economy in ASEAN Edited by Lurong Chen and Fukunari Kimura COVID-19 in Indonesia Impacts on the Economy and Ways to Recovery Edited by Lili Yan Ing and M. Chatib Basri Robots and AI A New Economic Era Edited by Lili Yan Ing and Gene M. Grossman Local Content Requirements Promises and Pitfalls Edited by Lili Yan Ing and Gene M. Grossman Routledge-ERIA Studies in Development Economics For more information about this series, please visit: www.routledge.com/Routledge-ERIAStudies-in-Development-Economics/book-series/ERIA Local Content Requirements Promises and Pitfalls Edited by Lili Yan Ing and Gene M. Grossman First published 2024 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 selection and editorial matter, Lili Yan Ing and Gene M. Grossman; individual chapters, the contributors The right of Lili Yan Ing and Gene M. Grossman to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. The Open Access version of this book, available at www.taylorfrancis. com, has been made available under a Creative Commons AttributionNon-Commercial-No Derivatives 4.0 license. Funded by Economic Research Institute for ASEAN and East Asia (ERIA). Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Ing, Lili Yan, editor. | Grossman, Gene M., editor. Title: Local content requirements : promises and pitfalls / edited by Lili Yan Ing and Gene M. Grossman. Description: Abingdon, Oxon ; New York, NY : Routledge, 2024. | Series: Routledge-ERIA studies in development economics | Includes bibliographical references and index. Identifiers: LCCN 2023029276 (print) | LCCN 2023029277 (ebook) | ISBN 9781032542232 (hardback) | ISBN 9781032542218 (paperback) | ISBN 9781003415794 (ebook) Subjects: LCSH: International trade. | Foreign trade regulation. | Commercial policy. | Economic development. Classification: LCC HF1379 .L625 2024 (print) | LCC HF1379 (ebook) | DDC 382—dc23/eng/20230629 LC record available at https://lccn.loc.gov/2023029276 LC ebook record available at https://lccn.loc.gov/2023029277 ISBN: 978-1-032-54223-2 (hbk) ISBN: 978-1-032-54221-8 (pbk) ISBN: 978-1-003-41579-4 (ebk) DOI: 10.4324/9781003415794 Typeset in Galliard by Apex CoVantage, LLC Protection for sale. Love is not Michelle and Han Na LYI Nick, Dina, Rob, Shari, and Sam GMG List of figures ix List of tables x List of contributors xiii Acknowledgements xiv 1 Introduction 1 GENE M. GROSSMAN AND LILI YAN ING 2 Localization measures: a global perspective 14 DOROTHEE FLAIG AND SUSAN F. STONE 3 Local content policies in the mining sector 48 JANE KORINEK AND PAULO DE SA 4 The unintended consequences of high regional content requirements 87 KEITH HEAD, THIERRY MAYER, AND MARC MELITZ 5 LCR Policies in China and their impacts on domestic value added in exports 114 KUN CAI AND ZHI WANG 6 Conformity of Indonesia’s LCRs with its trade and investment commitments 145 MICHELLE LIMENTA, LILI YAN ING, JUNIANTO JAMES LOSARI, AND OSCAR FERNANDO Contents We specially thank Chatib Basri, Ernawati Munadi, Siwage Dharma, Dani Rodrik, Justin Yifu Lin, Miaojie Yu, Elhanan Helpman, Hal Hill, Tetsuya Watanabe, Shujiro Urata, Haryo Aswicahyono, Kiki Verico, Fauziah Zen, Iman Pambagyo, Mari Pangestu, and colleagues at the Ministry of Trade of Indonesia, the Ministry of Finance of Indonesia, CEPR, OECD, University of Indonesia, CSIS, Gadjah Mada University, Peking University, Sun Yat Sen University, Liaoning University, and University of Pelita Harapan for sharing their insights on LCRs. Ivana Markus and Livia Nazara provided excellent research assistance and Catherin Safitri provided generous administration support. We also thank Gilbert Gnanarathinam, Kendrick Loo, and Chelsea Low from Routledge, and Fadriani Trianingsih from ERIA. Acknowledgements DOI: 10.4324/9781003415794-1 This chapter has been made available under a CC-BY-NC-ND 4.0 license. Local content requirements in theory and practice Local content requirements (LCRs) have been used by many countries, both developed and developing, to promote the use of local inputs and support the growth of domestic industries. Initially, the term LCR (or, equivalently, “content protection”) was used to refer to a mandate that a certain fraction of domestically produced inputs, by value or by volume, be incorporated in any final good sold in the domestic market. Over time, the range of policies covered by the term has expanded alongside the increased range of localization practices used by various national and local governments. Now, outcomes may be legally mandated or aspirational. The outcomes may reference input shares, employment, firm-ownership shares, location of R&D, or technology transfer. LCRs may include restrictions on the provision of certain services, eligibility for government contracts, local performance of compliance tests, or the location of data storage. Aspirational targets might be incentivized with subsidized export or investment financing, with tax breaks, with price concessions for government-supplied energy or infrastructure, with conditional bailouts, or with other financial inducements. In this book, we use the term LCR broadly to include any laws or regulations that require or encourages the use of locally produced inputs or services in a multi-stage production process. LCRs also play a role in bilateral and regional trade agreements, where they are known as “rules of origin” (RoOs). Trade agreements generally call for preferential tariff treatment of goods emanating from a partner country. But such agreements must define what it means to “emanate from,” or else goods imported from outside the member countries may enter the region in a low-tariff country and then be shipped on to a high-tariff country after the addition of only minimal or negligible local value added. RoOs specify what fraction of the value added of an internally traded good must originate within the region in order that the good qualify for preferential treatment. RoOs might also further stipulate minimum percentages from each of the various countries within the region, as with certain provisions of the United States-Mexico-Canada Trade Agreement (USMCA). While perhaps originally 1 Introduction Gene M. Grossman and Lili Yan Ing 2 Gene M. Grossman and Lili Yan Ing intended to thwart transshipment, RoOs are regularly used now to encourage regional production. LCRs first entered the arsenal of trade instruments in Australia, which, in 1948, restricted the use of imported car parts in the local assembly operations of British multinationals while offering concessionary financing based on the fraction of Australian value added to encourage the production of “Australia’s own car” (Pursell, 2001). Several countries quickly followed suit, including Canada, which instituted LCRs to shield domestic parts producers from American competition prior to the conclusion of the Canada-American Automotive Agreement in 1965 (Wonnacott and Wonnacott, 1967; Johnson, 1971). LCR policies to foster import substitution in the automobile industry soon became commonplace in Latin America, where they were introduced in Chile, Argentina, Mexico, and Brazil (Munk, 1969; Johnson, 1967). Moreover, Australia quickly extended its use of this new instrument well beyond the automobile sector, implementing policies to encourage or require use of local inputs in industries as disparate as petrochemicals, tobacco, peanut oil, coffee, fruit juices, industrial machinery, and agricultural tractors (Lloyd, 1971). Still, LCRs were relatively uncommon when Corden (1971) and Grossman (1981) first analyzed their economic consequences. Their popularity waned in the 1980s as more and more countries became disillusioned with using a strategy of import substitution to promote development. LCRs have made a roaring comeback, particularly after the Global Financial Crisis in 2008. Between 2008 and 2013, almost 200 new LCR measures were introduced, according to the Global Trade Alert. This figure grew to more than 500 measures that were put into place during the period from 2014 to 2020. Moreover, the implementation of these policies has been widespread, ranging across developed and developing economies. Quite prominently, the United States has made LCRs a cornerstone of its recent policy to promote the development of electric vehicles as part of the Inflation Reduction Act of 2022. The range of economic activities targeted by LCR policies around the globe has expanded to include many resource-extracting sectors, information technology, healthcare goods and services, financial services, agricultural products, and others. Why do countries adopt LCRs that favor local sourcing of intermediate inputs and services? The list of arguments to support such policies mirrors those offered for protectionist trade policies more generally. First, LCRs might afford new job opportunities in certain sectors or regions of the economy. These jobs, in turn, might boost wages, reduce unemployment, or encourage investments in human capital. Second, LCRs, by encouraging specific local activities, might provide spillover benefits to other activities and sectors via research and development or learning by doing. Third, localization policies might encourage or mandate greater foreign direct investment, joint-venture partnerships, or technology transfer on terms favorable to the host country or at the expense of alternative hosts. LCRs applied to primary products are often Introduction 3 intended to encourage higher value-added activities, with the aim of promoting indigenous management skills and technological knowhow. Of course, LCRs, like other forms of protection, often fail to achieve these lofty goals. Whether introduced by well-meaning leaders or in response to special-interest lobbying, such policies often fail to generate positive spillovers of sufficient magnitude to justify the higher costs of domestic sourcing. The desired jobs may not materialize due to inadequate management, lack of requisite skills, unavailability of complementary inputs, or other reasons. Even if new job opportunities are generated in targeted sectors, they may come at the expense of employment in other sectors that potentially offer greater economic benefit. In short, policies that discriminate in favor of local producers of inputs and services may be subject to the same, unfavorable cost-benefit analysis as with other forms of trade protection. Numerous tales of disadvantageous LCRs are told in Hufbauer et al. (2013), Stone et al. (2015), and elsewhere in literature. Similar critiques of RoOs in bilateral and regional trade agreements may be found in Cadot et al. (2006), Krueger (2012), Conconi et al. (2018), Cadot and Ing (2019), and elsewhere. The theoretical literature, beginning with Grossman (1981) and Dixit and Grossman (1982), has identified a particular risk associated with LCRs that distinguishes these policies from tariffs and other forms of protection for domestic industries. Whereas tariffs on final goods boost local demand for the protected goods and thereby demand for all inputs (including those produced locally) used in the production of these goods, LCRs that raise the cost of inputs can easily have unintended consequences. Alongside the import substitution mandated or encouraged by these policies comes an adverse “output effect”; as costs for downstream producers rise, these firms likely will scale back production and reduce their demand for inputs in the process. The offsetting substitution and output effects of LCRs may help to explain why empirical studies often find disappointing or even adverse effects of these policies on employment, value added, and foreign investment in targeted industries. Our book This book is intended to update the literature on local content requirements and rules of origin and to further understanding of the experience with and consequences of such policies and their consistency or not with the rules established by the world trading system. The distinctive features of our book are twofold. First, the research reported here uses the most up-to-date catalogs of LCRs that have been growing rapidly since the Global Financial Crisis of 2008. Second, the research complements analysis of LCRs at the global level with countryand firm-specific exercises. The remainder of this book contains seven chapters that can loosely be divided into three parts. The first part, comprising Chapters 2 and 3, provides an overview of major LCR policies used globally (Chapter 2) and in mineral exporting countries (Chapter 3). The second part focuses on important 4 Gene M. Grossman and Lili Yan Ing recent LCR policies in the world’s two largest economies, the United States and China. Chapter 4 concerns the effects on the organization of the North American automobile sector of the new RoOs in the USMCA while touching also on the implications for the car industry of Britain’s exit from the European Union. Chapter 5 addresses China’s industrial policy initiatives such as “Made in China 2025” that are intended to promote further industrialization and innovation in that country. The final three chapters shine a spotlight on Indonesia, a large, emerging economy that is a major exporter of natural resources. Indonesia is interesting for our purposes because its laws and regulations include a variety of LCR policies and because it provides a test case for the consistency of such policies with the rules agreed by members of the World Trade Organization (WTO). We proceed now to describe the contents and main contributions of these chapters in somewhat greater detail. Chapter 2 by Dorothee Flaig and Susan F. Stone discusses the recent experience with LCRs in the world economy. They begin by reviewing the reasons why countries implement policies that stipulate or incentivize the use of domestic inputs in local production. Among the most prominent objectives that they cite are employment objectives and technology transfer. Next, the authors discuss trends in the implementation of LCRs, pointing to an acceleration of usage from the period of 2008–2013 to the more recent period of 2014–2020, as reported by Global Trade Alert. The authors cite India, Germany, and the United Kingdom as the most intensive users of LCRs, but they qualify this observation by pointing out that counts of usage do not account for heterogenous impact and that the Global Trade Alert tallies LCRs in multiple jurisdictions within a country, so that more decentralized polities are likely to have higher counts. The heart of the chapter uses the OECD Trade Model, METRO, to provide a quantitative evaluation of seven major instances of new LCR policies, chosen from a sample of 565 measures that were considered for the purpose. The measures under review were selected as those that could be modeled quantitatively and that were likely to be among the most trade distorting. In order to apply the METRO model, a policy must specify an identifiable sector and an identifiable restriction that could be meaningfully enforced, it must affect a sufficiently large sector or region of the economy, and it must be a binding measure applied where the domestic sector has capacity to meet the required demand. Application of these criteria yields a good sample of the types of LCR measures that have been applied recently and of the types of economies that have applied them. Specifically, the authors focus on (i) tax credits available to Argentinian car producers that use specified percentages of local content; (ii) a requirement imposed by Brazil on the telecommunications sector that they use a minimum of local content in their 4G networks; (iii) the preference margins allowed by the Brazilian government for public procurement of a variety of nationally produced goods; (iv) regulations introduced by the Indonesian government that essentially require the entire assembly process Introduction 5 of motor vehicles and motorcycles to take place locally; (v) a regulation that makes it mandatory for Saudi Arabian governmental agencies to purchase their medical supplies from the local industry; (vi) a Mining Charter in South Africa that establishes a minimum local content requirement for mining goods and for total services used by firms in the domestic mining industry; and (vii) the Buy America program that is required of U.S. states that receive grants for transportation funding from the federal government. METRO is a static, global, computable general equilibrium (CGE) model. It incorporates many countries, sectors, and factors of production and distinguishes output from each sector according to its end use. The authors model each of the LCR measures as a restriction of calibrated magnitude on the input choices of sectors in the affected countries. After solving the baseline model under the assumption that the LCR policies are not binding, they resolve the model imposing their calibrated restrictions. The simulations provide estimates of the effects of the seven policies on real GDP, trade flows, labor income, total disposable income, and the terms of trade. They also generate disaggregated estimates of the effects of the policies on imports and production in the 27 sectors captured by the model. Flaig and Stone estimate that the economic impacts of the seven LCR measures they study are modest but generally negative. The LCRs tend to undermine the long-run competitiveness of the sectors in which they are applied while having limited or negligible spillover effects on the broader economies. Since the model assumes full employment, where the policy targets a large sector – such as the LCRs in the automobile industries in Argentina and Indonesia – the expanded use of local inputs necessarily comes at the expense of other sectors of the economy. As the non-targeted industries substitute away from domestic inputs, their demand for imports grows, with a potentially negative (albeit small) effect on the terms of trade. Chapter 3 by Jane Korinek and Paolo De Sa builds on Korinek and Ramdoo (2017) and focuses on laws on regulations that seek to stimulate growth of local industries upstream and downstream from the mining sectors in resource-rich economies. The chapter begins with a typology of LCRs that distinguishes mandatory measures from incentives-based policies and supplyside policies from demand-side policies. Demand-side policies intended to promote backward linkages between mining firms and their suppliers include preferences for local suppliers of inputs used in mining and numerical targets for employment by extractive firms. Supply-side policies with upstream suppliers include requirements to provide training, fund capacity development, and conduct awareness campaigns about procurement opportunities. Demandside policies aimed at increasing interactions with firms and industries downstream from the mining sector include restrictions or taxes on mineral exports beyond those on processed products, requirements that extractive firms sell a specified share of their output domestically, or tax concessions that favor local sales. On the supply side, LCRs might require extractive firms to invest in downstream processing facilities or to collaborate with training institutions to 6 Gene M. Grossman and Lili Yan Ing promote the development of needed skills. Incentive-based policies to encourage forward linkages might include tax concessions based on domestic sales of mined materials or subsidized loans for capacity investments. The authors note that mandates have been more popular in African countries such as Ghana, South Africa, Tanzania, and Zambia, whereas more developed economies such as Australia, Canada, and Chile have relied more heavily on an incentive-based approach. Korinek and De Sa go on to highlight a number of common reasons LCRs in the mining industries have generated disappointing results. First, an inadequate appreciation of mining firms’ inputs needs and of the absorptive capacity of local suppliers has led many countries to set unrealistic targets for local procurement that domestic firms have been unable to satisfy. Second, some countries have employed broad definitions of local content, which provides flexibility to the sector but makes it difficult to assess gains in value added and spillovers to the rest of the economy. The available evidence suggests that mandatory, quantitative LCRs have failed to generate significant growth in the use of locally sourced inputs by the mining sector, nor have they strengthened linkages with upstream industries. Meanwhile, requirements or incentives for downstream processing have failed in the long run when processing firms have not been able to produce goods of sufficient quality or to achieve international cost competitiveness. Finally, the authors note, LCRs have contributed to government deficits even when they appear on paper to be fiscally neutral, because they decrease profitability in the mining sectors and thereby reduce the governments’ receipts from corporate taxes and royalty payments. The authors conclude that when used at all, LCRs should be part of a comprehensive policy to promote institutional development and to foster intersectoral partnerships. They note that government support for sectors that rely on local mineral extraction may exacerbate the impacts of materials price fluctuations, harm the local ecology, and interfere with diversification of the broader economy. Instead, Korinek and De Sa suggest that resource-rich countries devote greater attention to promoting macroeconomic stability, removing barriers to entry, providing a transparent and stable regulatory environment, and improving local infrastructure, institutions, and skill levels. Chapter 4 by Keith Head, Thierry Mayer, and Marc Melitz analyzes the tighter rules of origin that now apply to the North American automobile industry following the renegotiation of the regional trade agreement that was formerly NAFTA and now is USMCA. NAFTA required regional content of 62.5% of value for cars to qualify for duty-free entry into one of its members. The USMCA raised this regional content requirement to 75% and introduced additional, binding requirements. The tightening of the RoOs and the adding of additional requirements clearly were intended to discourage firms from sourcing parts from outside North America, which the parties (and especially the United States) hoped would bolster demand for North American parts. In a companion paper, Head et al. (2022), the authors extend the oneinput Grossman (1981) model to include many inputs. However, their main Introduction 7 result, unlike what was emphasized by Grossman, does not rely on a reduction in the number of completed automobiles. Instead, they make another important observation. Firms in the automobile industry retain the option to choose the cost-minimizing source of parts, provided they are willing to sacrifice the treaty’s tariff benefit and pay the MFN tariff whenever a completed car crosses a border. Indeed, Head et al. document in Chapter 4 that compliance with the RoOs has declined since the introduction of the stricter USMCA rules. If compliance rates fall sufficiently, a tighter RoO intended to expand value added within the region might have the opposite effect. The authors term the inverted-U shape relationship between the strictness of RoOs and the regional value added a “Laffer curve for RoOs.” The authors take their model to the data, using detailed information on the source of engines and transmissions for all car models assembled in North America. Since the U.S. MFN tariff on automobiles is only 2.5% and only a fraction of the cars assembled in Mexico are exported (a fraction that varies by model), it stands to reason that some producers will be willing to pay the tariff in lieu of sourcing more expensive parts. Simulation of the model predicts that 16.9% of models that complied with a binding RoO under NAFTA will become non-compliant under the tighter RoOs of USMCA. These firms account for a predicted 11.9% fall in employment in plants that manufacture parts, offsetting the 23.6% rise in employment that the model predicts for plants that choose to comply with a binding RoO ex post. Overall, the model predicts employment gains of only 2.3% in plants that manufacture parts, much smaller than the 20% gains that would have been expected had they assumed that all carlines comply with the new RoOs. Had the new RoOs been set at 85%, as the United States had initially demanded, employment in parts manufacturing actually would have declined, according to the model’s estimates. The authors also simulate the effects of BREXIT on the European car industry. After Britain’s withdrawal from the European Union, the country negotiated a new pact known as the European Union and United Kingdom Trade and Cooperation Act. The TCA requires regional content of 55% for British cars to enter duty free into the European Union and for EU cars to enter similarly into Great Britain. The authors note that 85% of carlines already satisfied this requirement prior to TCA, suggesting that the plants manufacturing these models will be little affected. On the other hand, the MFN tariff in both Britain and the European Union is 10%, much higher than the level in the United States, suggesting that firms that did not already satisfy the new RoOs beforehand may well choose not to comply. The simulations again indicate that a fall in employment in parts manufacturers serving firms that choose not to comply will offset the employment gains in firms that comply with a newly binding constraint. Overall, the gains in employment are predicted to be less than 1%. The analysis by Head et al. in Chapter 4 emphasizes that overly strict LCRs that are not subject to mandate but rather are supported by fiscal incentives may be counterproductive. Firms that would have complied with a milder restriction may opt out once the requirements become too severe. 8 Gene M. Grossman and Lili Yan Ing Chapter 5 by Kun Cai and Zhi Wang analyzes LCR policies in China, with a particular emphasis on the Made in China 2025 policy. The chapter begins with an overview of China’s LCR framework, noting that these policies were explicit prior to China’s accession to the WTO but became more opaque afterward. Although the legally mandated LCR percentages for goods or services were gradually lifted, implicit localization biases ingrained in the implementation of industrial policies took their place. On the surface, these policies treat producers similarly regardless of nationality, but in practice, only indigenous firms can benefit from many of the preferential policies, and when foreign producers are able to participate, often they are “encouraged” to transfer technology or to source locally. The opaque nature of the current LCR policies makes them difficult to measure precisely, but the authors argue that their effects are very heterogeneous across sectors. Moreover, Beijing has launched a recent campaign to encourage the development of more advanced technologies at home to rely less on the United States and other Western suppliers. China aims to bolster indigenous firms’ capacity for innovation and to have them become global leaders in advanced technologies. To further these objectives, various government agencies in China at different levels have implemented a series of industrial policies, including some implicit LCRs that benefit local firms. Cai and Wang are especially interested in the impacts of China’s LCR policies on the domestic value added embodied in exports. They extend the Koopman et al. (2012) methodology for using world input-output matrices to attribute value added in exports to source countries so as to allow for circumstances as in China, where a sizable fraction of exported goods is produced in export processing facilities that are able to import their raw materials and components duty free. The authors use the input-output matrices published by China’s National Bureau of Statistics for 2007, 2012, and 2017, along with detailed trade statistics from China Customs. Using the changes that occurred during the periods between the publication of these data, they estimate the impact of China’s LCR policies on the fraction of domestic value added in total exports, manufacturing exports, and exports of foreign-owned firms. In the aggregate, domestic value added in exports rose from 64.6% in 2007 to 65.3% in 2012 and to 69.9% in 2017. The estimated gains are smaller when only manufacturing is considered, and they mask opposing trends for normal and processing exports. Whereas the domestic value-added share in normal manufacturing exports increased from 2007 to 2017, the share in processing exports actually declined. Focusing only on foreign owned firms, they find a slight decline over the period. The authors also study changes in domestic value added in exports at the industry level, comparing only 2012 with 2017, because industry definitions changed after 2007. Roughly one-third of the 68 industries, accounting for 20% of total exports, had a domestic value-added share in exports between 51% and 75% in 2012. These were mostly capital-intensive industries such as basic chemicals, iron and steel, lifting and handling equipment, pumps, Introduction 9 generators, and batteries. Roughly half of the industries, accounting for 25% of total exports, had a domestic value share in the earlier year greater than 75%. These industries were more labor intensive, including textiles, apparel, footwear, leather, and furniture. The lowest domestic value-added shares were found in the more technologically sophisticated industries, such as computers, communication equipment, and electronic components, where foreign ownership plays a major role. Comparing the estimates for 2012 and 2017, 15 of the 68 industries saw their domestic value-added share in exports rise by more than 5%, and 4 experienced an increase of greater than 9%. Cai and Wang concede that their methodology does not allow them to test for a causal relationship between LCR policies and domestic content outcomes. Still, their accounting decomposition leads them to conclude that the various LCR policies implicit in China’s industrial strategy did not seem to play a significant role in promoting increased local content in China’s exports between 2007 and 2017. Chapter 6 by Michelle Limenta, Lili Yan Ing, Junianto James Losari, and Oscar Fernando is the first of three that focuses on LCRs in Indonesia. Indonesia is among the countries with the highest incidence of local content restrictions, with LCR policies dating back to the 1950s. LCRs have long appealed to Indonesian policymakers who are keen to encourage domestic value added and expand employment in the industrial sector. Limenta et al. consider in detail whether Indonesia’s LCR policies are consistent with its commitments under its various multilateral and regional trade and investment agreements. They begin by outlining the justifications offered by the Indonesian government for its various policies. These include the country’s goal of upgrading competitiveness and eliminating its trade deficit, which the government believes can be achieved by increasing local value added in traded sectors. The use of LCRs began in Indonesia with the “Benteng Program” (1950– 1957), followed by the “Deletion Program” (1974–1993) and the “National Car Program” (1996). After a period of dormancy, Indonesia resumed in 2009 a local content strategy with the implementation of the “Increased Use of Domestic Production” program. This policy introduced local content requirements for goods and services purchased by the government. In 2013, Indonesia implemented the aforementioned regulation stipulating a minimum percentage of domestic oil and gas as input into the local production of many goods. Since that time, many members of the WTO have challenged Indonesia’s policies in the Trade-Related Investment Measures (TRIMS) Committee, claiming that they run counter to commitments that Indonesia made to its trade and investment partners. Of particular concern to the WTO members have been policies regulating the use of domestically produced energy, as well as policies addressed to the telecommunications, pharmaceutical, and retail sectors. The authors proceed to review Indonesia’s relevant commitments under TRIMS, the General Agreement on Tariffs and Trade (GATT), the General 16 Dorothee Flaig and Susan F. Stone competition for the target industry and lead to a deterioration in product quality, as they reduce access to technologically advanced inputs and provide little incentive for internal innovation (Hufbauer et al., 2013). Corruption and favoritism from opaque and ad hoc policy design can also increase the longrun negative impact of these policies (Kuntze and Moerenhout, 2013; Weiss, 2016). The objectives of LCRs – such as building up a competitive industry through stronger industrial links, creating new suppliers and backward linkages – is rarely obtained (Hufbauer et al., 2013). In most cases, LCRs isolate high-cost producers from global competition and innovation and result in insufficient incentives for research and development (R&D) investments. Most LCRs implemented have employment as their primary objective, explicitly or implicitly stated. The use of domestic suppliers has an immediate job effect that can be particularly powerful during economic downturns. LCRs with employment objectives encompass goals such as creating new jobs, creating higher-skilled jobs, and increasing national income. However, these policies often sacrifice job growth in the general economy for job growth in the targeted sector. The United States (US) Buy American Act, 1933, is estimated to have cost about 360,000 jobs in non-targeted sectors throughout the US economy (Dixon, Rimmer, and Waschik, 2018). Measures targeting technological development tend to require foreign firms to transfer technology to domestic operations or domestic suppliers. To the implementing economy, this technology transfer is seen as an efficient way of increasing competitiveness in world markets. The specific goals of technological development can include improving technological capacity and spurring innovation at the national, regional, or industry level. However, these goals are often undermined by a lack of available skills. Factors that impact the ability to meet an LCR’s stated policy objective can be characterized across four areas: (i) market size and stability, (ii) policy design and coherence, (iii) the restrictiveness of the LCRs, and (iv) the domestic industrial base (Kaziboni and Stern, 2021). If the domestic market is small or unstable and cannot meet the demands of the local producers, there is a risk that these local producers exit the market, defeating any employment or technology-transfer goals. A policy that is too vague is usually unenforceable, and it will not be effective. Moreover, if the LCR is set at a level that does not have a meaningful impact on the sourcing decisions of the importer, no change will occur. Some LCRs are set below existing sourcing levels, having no impact on market decisions. Finally, if there are no credible producers of the input, then any LCR policy will simply lead to firms exiting the market. Governments often attempt to achieve several policy objectives with one LCR (e.g., increase output and employment along with technology transfer). In these instances, the policy often ends up having contradictory outcomes (e.g., increasing production but decreasing productivity if labor is unable to implement the transferred technology). Subsequent productivity declines, coupled with shortages of sufficiently skilled labor, may lead to an overall lower level of labor demand, showing that one policy is often unable Localization measures 17 to hit two targets (Fang, 2020). Policymakers often neglect to identify practical challenges that might negatively impact the efficiency and effectiveness of LCRs when adopting these measures. There is limited consideration of the fact that the economic impact of LCRs is complex and depends on several variables, including their interaction across policy areas (Lin and Weng, 2020). The political influence of producers can also affect the level of local participation. Ablo (2017) shows that while LCRs have the potential to promote links between some sectors and the rest of the economy, the degree to which producers support the government may limit the extent to which significant local content can be achieved. Thus, the political relationship between the government and industry, the capacity of local SMEs, and the techniques and practices of multinational companies will all have a bearing on the effectiveness of any LCR policy. By limiting competition and input choices, the target firm faces a limited, if not single, supplier – leading to higher input costs along the production line. This can also affect the quality of the material/input, which can further impact costs. These higher costs are then passed on, in whole or in part, downstream, increasing costs to both the consumer and producer. This ultimately means higher prices for the end user. Another way LCRs can have a detrimental effect on the domestic market is through resource allocation effects. Resources being shifted to the targeted firm/industry become scarcer elsewhere in the domestic market, raising costs to other firms. Another longer-term spillover in the domestic economy relates to innovation and skills development. While evidence shows that targeted firms will undertake training and development of local firms to meet LCR targets (Ramdoo, 2015), this leaves little motivation in the domestic firms to innovate, as they have little or no competition to spur such innovation. This can affect the degree to which firms across the economy engage in innovative behavior. This lack of innovation can create the longest-lasting, most detrimental effects of an LCR. A recent study (Kingiri and Okemwa, 2022) shows that local content policies have not had a positive impact on technology development in the Kenyan renewable energy sector. From the perspective of public expenditures, LCRs reduce import tariff revenue as well as potential corporate tax revenue by increasing the operating costs and reducing the profitability of multinational companies (Kolstad and Kinyondo, 2017). Imposing LCRs has the opportunity cost of forgone taxes, which could be used in more effective ways to improve development prospects. In the case of incentive-based LCRs, governments often forego revenue from or provide incentive payments to these ventures, which has direct public expenditure implications. In addition, past experiences of resource-rich developing countries indicate that local content policies can exacerbate key problems of patronage and rent-seeking, increasing the danger that the resource wealth will undermine rather than help development (Kolstad and Kinyondo, 2017). This chapter extends and updates our earlier study from 2015 (Stone, Messent, and Flaig, 2015). It discusses recent instances of LCRs and, by modeling 18 Dorothee Flaig and Susan F. Stone representative examples of the policy, provides some insights on the impact they have on the economy. Section 2 examines recent trends in their implementation. Section 3 outlines the measures examined for this study and the modeling approach adopted. Section 4 provides the simulation results, while Section 5 presents some concluding thoughts. 2 Recent LCR implementation The use of LCRs has been accelerating in recent years. According to Global Trade Alert,1 countries put in place more than 500 individual local content measures from 2014 to 2020 compared with less than 200 measures from 2008 to 2013 – a 155% increase (Figure 2.1). Not only has the number of measures increased, the way in which they have been implemented has changed. The less transparent types of instruments have risen, with the Global Trade Alert measures rated amber (“likely” to cause discrimination) increasing significantly over those rated red (“almost certainly” causing discrimination against foreign companies). That is, the number of measures that clearly state the type and requirements of an LCR restriction has declined as a share of the total number of LCRs imposed. Only 6% of the measures were rated amber during 2008–2013, whereas 32% of the measures were rated amber during 2014–2020. On the surface, India, Germany, and the United Kingdom appear to be the main users of LCRs (Figure 2.1). However, the actual impact of these measures is much more complex. First, as noted, Global Trade Alert provides a count of the incidence of a measure, not its impact. Thus, a measure affecting a specific small sector (e.g., Argentina’s LCR on certain types of medicines from Spain) counts the same as a general measure affecting a much larger sector (e.g., Russian restrictions on auto parts to the automotive sector for all trade partners). Second, a tightly binding measure counts the same as one that is only partially binding or not binding at all. In addition, many economies implement these measures at a variety of jurisdictional levels. Larger economies with decentralized economies (e.g., the US and India) have many measures implemented at the level of the individual state or even the local level of government. Others have a much more centralized approach, while still others (e.g., France and Germany) have policies implemented at the supranational level. The variety of instances makes identifying and measuring LCRs challenging. Global Trade Alert reports LCRs across four intervention types: labor, operations, sourcing, and incentivizing. Labor generally refers to LCRs tied directly to hiring and employment requirements. Operations are LCRs that have requirements concerning a firm’s ability/permission to operate in the domestic market. Sourcing refers to the requirements to source inputs (parts and components) from local manufacturers. Incentives are tax or other government benefits received when buying or using local inputs, operations, or labor. Localization measures 19 020406080100 120140 160 United States United Kingdom Turkey Saudi Arabia Russia Indonesia India Germany China Brazil Australia Argentina Total 2008-2013 Total 2014-2020 Figure 2.1 Incidence of Local Content Requirements, 2008–2020* (Selected Economies) BNDES = The Brazilian Development Bank, LCR = local content requirement. Note: The numbers represent individual instances of LCRs, so their actual impact is not directly comparable. For example, in 2015, BNDES financed three wind parks with $260 million. As these measures affected different trading partners and different sectors within wind turbines, they accounted for 57 of the 594 measures of local sourcing for Brazil. Argentina’s LCR on mining counts as one measure yet affects more than 10 sectors measured at the 2-digit level across all trading partners. * The large number of LCRs attributed to Germany all relate to support given under Germany’s Export Credits program. Source: Global Trade Alert (2022). Figure 2.2 shows how the types of LCRs used by governments have changed over time. Whereas in the period right after the Global Financial Crisis (2008– 2013) LCRs focused on ensuring that inputs and labor were sourced locally, the later period (2014–2020) switched to incentivizing firms by offering tax breaks, preferential lending, or other perks tied to using local inputs or establishing local production. The number of measures offering incentives more than doubled from less than 6% of all measures in 2008–2013 to more than 12% in 2014–2020. As described by Deringer et al. (2018), Global Trade Alert ranked tracked policy interventions by their possible damage to foreign trade and investment. According to this ranking, LCRs representing public procurement localization are ranked fifth, and measures representing other localization requirements are ranked seventh (Evenett and Fritz, 2021). Ranking ahead of LCRs are (i) state aid, (ii) trade defense, (iii) import tariffs, and (iv) export taxes or restrictions, with trade finance measures in sixth place (Evenett and Fritz, 2021). LCRs 20 Dorothee Flaig and Susan F. Stone are also frequently implemented in the form of discriminatory government procurement. These measures reduce the number of eligible firms allowed to compete in a market and thus decrease output and employment while increasing procurement costs and market power (OECD, 2020). Measures related to data localization are among the fastest-growing types of LCR measures. Such measures attempt to control the movement of personal data and local storage and processing of data (López González, Casalini, and Porras, 2022). As data flow is becoming an essential aspect of trade, related LCR measures affect most sectors within an economy (OECD, 2020). Some experts perceive this type of protectionism as “perhaps today’s greatest threat to the further liberalization of the global trading system” (Ezell, Atkinson, and Wein, 2013).2 At the same time, the increasing connection of trade and data flows may also provide a stronger rationale for such measures to ensure privacy and identity security. 3 Defining measures to be addressed It has been argued that localization barriers add to the cost of doing business domestically and internationally, leading to a distortion of world trade flows and lost market opportunities. However, the studies attempting to quantify these impacts across global markets have been limited. This chapter will update one such attempt (Stone, Messent, and Flaig, 2015) by estimating the impact LCRs implemented from 2008 to 2013 Labor Sourcing Operations Incentivizing LCRs implemented from 2014 to 2020 LaborSourcing Operations Incentivizing Figure 2.2 Type of Local Content Requirements LCR = local content requirement. Source: Global Trade Alert (2022). Localization measures 21 of a set of LCRs on international trade, using the Organisation for Economic Co-operation and Development (OECD) trade model, METRO. This set of LCRs is defined from information taken from several data sources and constitutes current in-force LCR policies that were put in place from 2014 to 2020.3 Similar to the work undertaken in Stone, Messent, and Flaig (2015), several sources were consulted. These include the following: • Peterson Institute for International Economics Local Content Requirements: A Global Problem (Hufbauer et al., 2013) • Global Trade Alert online database • European Commission (2022) Market Access Database • World Bank (2022) Temporary Trade Barriers Database • World Trade Organization (2022) Trade Monitoring Database More than 565 measures were considered for the study. All the identified LCRs were then reviewed to assess their affinity to quantification. The quantitative analysis presented here focuses on measures that tend to be the most trade distorting. These are measures that restrict access to markets and measures that render price preferences tied to a specific level of domestic content. Input measures that determine market access accounted for most of the measures examined for this report. To arrive at a list of measures whose impacts could be quantified, several criteria had to be met. Following the discussion on localization characteristics already, four characteristics can be identified: 1 Identifiable sector – many of the LCRs were broad statements about “supporting” domestic sourcing without direct reference to a particular sector or region of economic activity. For example, countries put in place LCRs for government procurement. These are blanket policies that may or may not be implemented in any specific sector. Given that there is no way to identify which, if any, sectors were affected by these policies, these measures were excluded from the analysis. 2 Identifiable restriction – if the restriction is not clearly indicated, it cannot be meaningfully enforced and thus cannot be modeled. For example, Turkey introduced localization requirements on remote programmable e-SIM technologies without stipulating the size of these requirements. 3 Sufficient size – if the sector or region is not significant, it will have little impact on the market. An example is an Argentine law obliging automobile fuel producers to use bioethanol from the northeast of the country. Neither the bioethanol market nor the northeast region is sufficiently large to be captured in an economy-wide model. 4 Enforceable restriction – there must be a domestic sector that can meet the required demand, and the restriction must be binding (i.e., the measure is excluded if the domestic content already meets or exceeds the level called for in the LCR). 22 Dorothee Flaig and Susan F. Stone Many of the instruments examined were implemented at the subnational level, which the model does not cover. As noted above, policies related to broad goals such as national security or government procurement can be applied over multiple sectors or not and are at the discretion of government agents. Finally, for many of the measures reported, we were unable to determine if they are still in force. Thus, these were also dropped from the analysis. Applying these criteria, we were able to identify seven LCRs that meet the conditions necessary for modeling. They provide a good sample of the types of LCR measures applied and the economies applying them. Both developed and developing countries are included in the analysis. This study estimates the economic impact of LCRs imposed in selected subsectors of automobiles, mining, medical supplies, telecommunications, and transport. Finally, some “buy local” procurement provisions are at such a level that noncompliance would undermine a firm’s competitive position, making them a “requirement.” Thus, we include examples of two such government procurement provisions. Argentine automotive sector. Since 2016, Argentine car producers can obtain a tax credit, allowing them to defer value-added tax (VAT). This tax credit is conditional on a minimum LCR in the final product. The minimum LCR is 30% for cars, trailers, engines, and agricultural vehicles; 25% for trucks; and 10% for automobile parts. The tax benefit is dependent on the level of local content and ranges from 4% to 15% of the sales value. The Argentine motor vehicle industry has an approximate sales value of $4.5 billion, so the value of this credit is $180 million to $675 million. Brazilian telecommunications sector. In 2017, Brazil implemented a requirement that the overall level of local content in the equipment used in its 4G networks must be at least 70%. The estimated size of the telecommunications sector in Brazil is more than $8.5 million, with almost 200 million 4G broadband subscribers. The estimated investment in telecommunications equipment was about $5.8 billion in 2018, implying a potential market of more than $4 billion solely available to domestic suppliers. This requirement comes on top of existing LCRs in the Brazilian telecommunications sector (Stone, Messent, and Flaig, 2015). Brazilian government procurement. In 2013 and 2014, the government of Brazil increased the preference margins for the public procurement of a variety of nationally produced goods. Preference margins are the maximum extent to which the price quoted by a local supplier may be above that of a competitor. The new margins range between 9% and 25% and include IT (15%), tractors (20%), airplanes (9%–25%), various ITrelated goods and services (15%–25%), capital goods (15%–20%), and toys (10%). These margins were deemed sufficiently high, given the various market sizes, for firms to be compelled to use local suppliers to remain competitive. Localization measures 23 Indonesian automotive industry. To promote the Indonesian motor vehicle components industry, several regulations implemented from 2014 to 2017 require de facto the entire assembly process of motor vehicles and motorcycles to take place locally. To have access to the Indonesian market, the LCR requires all major vehicle components and related services to take place within Indonesia. The Indonesian automotive market accounts for about 10% of Indonesia’s gross domestic product (GDP), or roughly $10 billion, with almost 25% destined for export markets. Saudi Arabian medical supplies. Since 2020, it is mandatory for government agencies to procure a range of medical supplies domestically, such as sterilizers, face masks, personal protective equipment for health practitioners, sterilization supplies for medical tools, and other medical supplies. The market for medical supplies in Saudi Arabia is estimated to be more than $2 billion. South African mining. The Mining Charter adopted in 2018 by South Africa establishes a minimum LCR of at least 70% for mining goods and 80% for total service expenditure in the sector. In addition, at least 21% of mining goods and 50% of services must be produced by a domestic company owned and controlled by “Historically Disadvantaged Persons,” another 5% (goods)/5%–15% (services) must be produced by companies owned by women or youth, and 44% (goods)/10% (services) by companies compliant with local Black Economic Empowerment (BEE). In 2018, the mining sector accounted for $22.5 billion of South African GDP and employed an estimated 456,000 workers. US Buy America. This far-reaching program has many provisions implemented on a preferential-treatment basis. However, transportation grants across many US states have a specific requirement for local content to be eligible to receive funding. Given that the main source of transportation funding is the federal government, for most US states, this amounts to an LCR. These programs stipulate that any public project funded by Transportation Investment Generating Economic Recovery (TIGER) grants must use some level of domestically produced iron, steel, and other manufactured goods. The amounts of both the grants and the LCR vary by individual states and projects within states but are estimated to be worth more than $4 billion. Model and data LCRs may have short-term benefits in terms of specific policy objectives, but adverse effects develop over time and often outweigh the short-term benefits. As with any model, not all the impacts of the policy will be fully reflected in the results. However, using a computable general equilibrium (CGE) model allows for the capture of these longer-run impacts, not to mention the effect these policies have on broader economic activity. The benefit of using a CGE 24 Dorothee Flaig and Susan F. Stone model in this analysis is its ability to capture impacts beyond the targeted sector, showing the effects these measures have on the rest of the economy as well as the global trade environment. The METRO model (Arriola et al., 2020) is based on empirical data and incorporates unique features of each region’s economic system. The model is calibrated to an augmented Social Accounting Matrix (SAM) version of the Global Trade Analysis Project (GTAP) version 10 database (Aguiar etal., 2019). The database features trade flows disaggregated by use categories derived from the OECD Trade in Value Added (TiVA) database as well as United Nations (UN) sources and bilateral remittance data from GTAP satellite data, i.e., GMIG2 (Walmsley, Winters, and Ahmed, 2007).4 These categories are intermediate use, use by households, use by government, and use by business/investment. The sector detail depends on the sector coverage in the GTAP database, which distinguishes 65 sectors with more detail in agri-food, and other sectors depicted largely on the International Standard Industrial Classification of All Economic Activities (ISIC) 2-digit level. This sector coverage does not allow modeling of LCR measures at a detailed sector level. The database is aggregated for this study, as detailed in Table 2.A1. The agriculture, food, and textile sectors in the GTAP database are aggregated; the study features 43 sectors, of which 4 are related to the extraction industries, 17 are related to manufacturing, and 20 are in the service sector. The database distinguishes eight factors of production, two skilled and three skilled labor types, capital, land, and natural resources. Countries are aggregated to larger regional composites, singling out the relevant countries of the selected policies examined. METRO is a comparative static global CGE model.5 Global CGE models link various markets, economies, and sectors – employing economic theory to show interlinkages between agents, sectors, and economies by simultaneously determining prices and quantities. The strength of METRO lies in the detailed trade structure and the differentiation of production and consumption commodities by use – intermediate, household, government, and capital consumption. The differentiation of commodity supply, and thus the resulting trade flows, by use category improves the ability to depict and analyze specific policy instruments such as LCRs. The remainder of this section gives an overview of key features of the model. Please refer to the METRO model documentation (Arriola et al., 2020) for a detailed and complete description of the model data and equations. The model is based on a series of regional SAMs, which are linked through trade relationships. Table 2.1 depicts the structure of the database, where income flows are read across rows and columns depict expenditures. Households, for example, receive income from factor services and remittance inflows and spend this income on private consumption, direct taxes, remittance outflows, and savings. Following accounting rules and depicting a complete and circular system, row and column sums must equalize. Thus, the total income ioral Relationshipse of the Database and BehavucturMETRO Model – Str Table 2.1 Expenditure Income flow Commodities (by sector, imported and domestic, by use category) Use category: Activities (by sector) Factors Use category: Household Use category: Government Use category: Capital Margins (bilateral, by use category) Rest of the world (bilateral, by use category) Commodities (by sector, imported and domestic, by use category) Use category: Activities (by sector) Factors (5 labor types, capital, land, and natural resources) Use category: Household Use category: Government Use category: Capital Margins (bilateral, by use category) Rest of the world (bilateral, by use category) Intermediate Private Public Investment: Margins exports: Exports: Threeinputs: consumption: consumption: Fixed shares Three-stage stage CET Leontief Stone-Geary Fixed shares CET functions functions input-output utility functions coefficients Domestic supply: Total supply from domestic production Value added: multilevel CES production functions Factor income: Remittance Fixed shares of inflows factor income Import tariffs, export Production Factor taxes: Direct taxes: taxes, sales taxes: Ad taxes: Ad Average tax Average tax rates valorem and specific valorem rates Depreciation: Private savings: Public savings: Current account Foreign savings: Shares of factor Shares of Residual balance on Current income household income margins trade account balance Trade and transport margins: Fixed coefficients Imports: Three-level Remittance CES outflows: Fixed proportion of disposable income mation., CET = constant elasticity of transfor felder (2013). CES = constant elasticity of substitution om McDonald and Thier: Adapted frceSour 32 Dorothee Flaig and Susan F. Stone centage Changesoduction by Sector, Perts and PrModel Results – Effects on Impor Table 2.3 Argentina– Brazil–telecom Indonesia–automobiles Saudi Arabia–GP South Africa–Mining automobiles medicals Imports Production Imports Production Imports Production Imports Production Imports Productio Agriculture 2.0 −1.8 0.0 0.0 0.9 −0.6 0.0 0.0 0.0* 0.0 Coal −0.2 −3.2 0.0 0.0 0.9 −1.4 0.0 −0.1 −0.1* 0.1 Oil 2.7 −1.3 0.0 0.0 1.1 −1.2 0.1 −0.1 0.0* 0.4 Gas 7.6 −5.4 0.1 −0.1 7.3 −6.1 0.2 0.1 −0.5* 1.4 Minerals 0.8 −0.7 0.0 0.0 0.5 −1.9 0.0 0.0 −0.8* −0.8 Food 3.6 −1.7 0.0 0.0 1.9 −0.8 0.0 0.0 0.0* 0.0 Textiles 6.0 −0.8 0.0 0.0 1.3 −2.9 0.0* −0.1 0.0* 0.0 Wood products 4.8 −0.2 0.0 0.0 3.8 −1.1 0.0* 0.0 0.0* −0.1 Paper production, 3.7 −0.9 0.0 0.0 1.1 −2.0 0.1* 0.0 0.0* 0.0 publishing Petroleum, coal products 1.0 −1.0 0.0 0.0 0.9 −0.2 0.1 0.0 0.0* 0.0 Chemical products 1.6 −2.7 0.0 0.0 1.2 −2.3 −2.0* 0.4 −0.3* 0.1 Basic pharmaceuticals 0.7* −3.7 0.0 0.0 2.6 −1.0 −2.3* 3.8 0.0* 0.0 Rubber and plastic 5.5 0.8 0.0 0.0 3.3 −1.0 0.0* 0.0 0.0* 0.0 Mineral products 4.0 0.1 0.0 0.0 2.9 −0.1 0.1 0.0 0.0* 0.0 Ferrous metals 7.6 1.3 0.1 0.0 3.5 −1.3 0.0 0.0 0.2* 0.1 Metals 6.8 −5.2 0.1 0.0 4.1 −4.2 0.0 −0.2 −0.1* −0.3 Metal products 8.4 5.7 0.1 0.0 3.2 0.4 0.1 0.0 0.1* 0.0 Computer, electronics 1.2* −0.8 0.0 0.0 0.3 −4.6 0.0 0.0 0.0* 0.1 Electrical equipment 1.4* 3.7 −0.7* 0.6 2.1 −4.8 0.0* −0.2 −0.1* 0.2 n Machinery, equipment 0.9* −4.2 −0.1* 0.1 1.4 −4.4 0.0 −0.1 −0.9* 0.6 Motor vehicles and parts −3.8* 39.2 0.0 0.0 −68.1* 43.3 0.0 0.0 0.0* 0.0 Transport equipment −9.7 17.1 0.0 0.0 −44.3* 23.3 0.0 −0.1 −0.9* 3.3 Manufactures 5.3 −0.2 0.0 0.0 3.1 −2.8 0.0 0.0 0.0* 0.0 Transport 1.3–1.8 −0.7–−0.1 0.0 0.0 1.3–2.1 0.1–3.6 0.0 0.0 −0.3–0.0 0.0–0.1 Communication 2.8 −0.3 0.0 0.0 1.5 −0.3 0.1 0.0 0.0 0.0 Utilities, construction 2.5–3.4 0.1–0.5 0.0 0.0 1.2–2.7 −0.5–0.0 0.0–0.1 0.0 −0.4–0.0 −0.1–0.3 Other service sectors 1.5–2.9 −0.6–0.2 0.0 0.0 1.0–2.0 −1.0–0.0 0.0–0.1 0.0 0.0 0.0–0.1 Note: The transport sector covers the three modes (air, land, and sea) and thus is reported as a range. * Sectors facing a binding local content requirement. Source: Model results. Localization measures 33 Table 2.3 Model Results – Effects on Imports and Production by Sector, Percentage Changes Argentina– automobiles Brazil–telecom Indonesia–automobiles Saudi Arabia–GP medicals South Africa–Mining Imports Production Imports Production Imports Production Imports Production Imports Production Agriculture 2.0 −1.8 0.0 0.0 0.9 −0.6 0.0 0.0 0.0* 0.0 Coal −0.2 −3.2 0.0 0.0 0.9 −1.4 0.0 −0.1 −0.1* 0.1 Oil 2.7 −1.3 0.0 0.0 1.1 −1.2 0.1 −0.1 0.0* 0.4 Gas 7.6 −5.4 0.1 −0.1 7.3 −6.1 0.2 0.1 −0.5* 1.4 Minerals 0.8 −0.7 0.0 0.0 0.5 −1.9 0.0 0.0 −0.8* −0.8 Food 3.6 −1.7 0.0 0.0 1.9 −0.8 0.0 0.0 0.0* 0.0 Textiles 6.0 −0.8 0.0 0.0 1.3 −2.9 0.0* −0.1 0.0* 0.0 Wood products 4.8 −0.2 0.0 0.0 3.8 −1.1 0.0* 0.0 0.0* −0.1 Paper production, publishing 3.7 −0.9 0.0 0.0 1.1 −2.0 0.1* 0.0 0.0* 0.0 Petroleum, coal products 1.0 −1.0 0.0 0.0 0.9 −0.2 0.1 0.0 0.0* 0.0 Chemical products 1.6 −2.7 0.0 0.0 1.2 −2.3 −2.0* 0.4 −0.3* 0.1 Basic pharmaceuticals 0.7* −3.7 0.0 0.0 2.6 −1.0 −2.3* 3.8 0.0* 0.0 Rubber and plastic 5.5 0.8 0.0 0.0 3.3 −1.0 0.0* 0.0 0.0* 0.0 Mineral products 4.0 0.1 0.0 0.0 2.9 −0.1 0.1 0.0 0.0* 0.0 Ferrous metals 7.6 1.3 0.1 0.0 3.5 −1.3 0.0 0.0 0.2* 0.1 Metals 6.8 −5.2 0.1 0.0 4.1 −4.2 0.0 −0.2 −0.1* −0.3 Metal products 8.4 5.7 0.1 0.0 3.2 0.4 0.1 0.0 0.1* 0.0 Computer, electronics 1.2* −0.8 0.0 0.0 0.3 −4.6 0.0 0.0 0.0* 0.1 Electrical equipment 1.4* 3.7 −0.7* 0.6 2.1 −4.8 0.0* −0.2 −0.1* 0.2 Machinery, equipment 0.9* −4.2 −0.1* 0.1 1.4 −4.4 0.0 −0.1 −0.9* 0.6 Motor vehicles and parts −3.8* 39.2 0.0 0.0 −68.1* 43.3 0.0 0.0 0.0* 0.0 Transport equipment −9.7 17.1 0.0 0.0 −44.3* 23.3 0.0 −0.1 −0.9* 3.3 Manufactures 5.3 −0.2 0.0 0.0 3.1 −2.8 0.0 0.0 0.0* 0.0 Transport 1.3–1.8 −0.7–−0.1 0.0 0.0 1.3–2.1 0.1–3.6 0.0 0.0 −0.3–0.0 0.0–0.1 Communication 2.8 −0.3 0.0 0.0 1.5 −0.3 0.1 0.0 0.0 0.0 Utilities, construction 2.5–3.4 0.1–0.5 0.0 0.0 1.2–2.7 −0.5–0.0 0.0–0.1 0.0 −0.4–0.0 −0.1–0.3 Other service sectors 1.5–2.9 −0.6–0.2 0.0 0.0 1.0–2.0 −1.0–0.0 0.0–0.1 0.0 0.0 0.0–0.1 Note: The transport sector covers the three modes (air, land, and sea) and thus is reported as a range. * Sectors facing a binding local content requirement. Source: Model results. 34 Dorothee Flaig and Susan F. Stone Table 2.4 Ar gentina – Motor Vehicle Sector LCR, Detailed Effects on Automotive Production, Percentage Changes (i) Motor vehicle sector Quantity Total Use category Intermediate inputs Private consumption Government consumption Prices Use category Capital goods Intermediate inputs Private consumption Government consumption Capital goods Production 39% 101% 17% 66% 28% −5% −5% −5% −4% Exports 51% 87% 31% 66% 38% −7% −3% −5% −4% Domestic 23% 175% 12% 43% 13% 0% −6% −8% −7% Imports −4% 3% −12% −2% −16% −1% −2% −1% −2% Total demand 7% 15% 5% 0% 0% −5% −2% −5% (ii) Transport equipment sector Quantity Prices Total Use category Use category Intermediate inputs Private consumption Government consumption Capital goods Intermediate inputs Private consumption Government consumption Capital goods Production 17% 11% 19% 84% 28% −4% −4% −4% −4% Exports 41% 32% 37% 84% 44% −3% −3% −4% −3% Domestic 14% 11% 17% 54% 23% −4% −5% −6% −5% Imports −10% −12% −10% −3% −9% −2% −2% −1% −1% Total demand 3% 6% 3% 0% −3% −1% −3% LCR = local content requirement. Source: Model results. Brazilian telecommunications industry The measure is modeled as a requirement to source 70% of the inputs of electronic equipment and machinery and equipment to the communications sector from domestic sources. Prior to the policy implementation, these sectors had local content of around 30%. The telecommunications sector is part of the post and telecommunications sector in the METRO model. Telecommunications Localization measures 35 account for about 80% in the broader sector (IBGE, 2022). To the extent that postal services use electronic and machinery equipment as an input, the impact of the LCR will be overstated. The model assumes that employed technologies do not change. Finally, to the extent that firms decide not to comply and leave the market, the impact of the LCR is likely understated. Results Telecommunications account for 9% of GDP, and 8% of the sector’s inputs are imported (Table 2.2). The LCR increases in domestic production in the targeted sectors are reported in Table 2.3 and Table 2.5. The model allows for the differentiation of domestic and export prices. This permits us to capture the ability of firms to engage in price discrimination, i.e., keeping their export prices low to protect or even increase market share overseas while raising domestic prices where, due to the LCR, competition is restricted. The electronic equipment sector benefits from increasing domestic demand due to the LCR exclusively from inputs into communications exports and production increasing 0.6% and 0.9%, respectively. This LCR builds on existing restrictive policies in the Brazilian telecommunications sector (Stone, Messent, and Flaig, 2015). Thus, the costs of this policy are limited and have no noticeable additional effects on other sectors or on the aggregate level of output or labor income (Table 2.3). However, what the model does not capture to a sufficient degree are the costs imposed on the government to monitor and enforce the policy. If these could be accurately quantified, it is likely that overall welfare impacts would be negative. Table 2.5 Brazil – LCR, Percentage Changes (i) Electrical equipment sector Intermediate inputs Private consumption Government consumption Capital goods Intermediate inputs Private consumption Government consumption Capital goods Quantity Prices Total Use category Use category Production 0.6% 3.1% 0.0% −0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Exports 0.9% 2.0% 0.0% −0.1% 0.0% −0.1% 0.0% 0.0% 0.0% Domestic 0.6% 3.6% 0.0% −0.1% 0.0% 0.1% 0.0% 0.0% 0.0% Imports −0.7% −1.3% 0.0% 0.0% 0.0% −0.1% 0.0% 0.0% 0.0% Total demand 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% (Continued ) 36 Dorothee Flaig and Susan F. Stone Table 2.5 (Continued) (ii) Machinery and equipment sector Quantity Total Use category Intermediate inputs Private consumption Prices Use category Government consumption Capital goods Intermediate inputs Private consumption Government consumption Capital goods Production 0.1% 0.5% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Exports 0.1% 0.3% 0.0% −0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Domestic 0.1% 0.6% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Imports −0.1% −0.2% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Total demand 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% LCR = local content requirement. Source: Model results. Brazil: preferential margins on various products in public procurementprocesses In 2013 and 2014, the government of Brazil decided to increase the preference margins for the public procurement of a variety of nationally produced goods. Preference margins are the maximum extent to which the price quoted by a local supplier may be above that of a competitor. The new margins range between 9% and 25% and include IT (15%), tractors with continuous tracks (20%), airplanes (9%–25%), various IT-related goods and services (15%–25%), capital goods (15%–20%), and toys (10%). The price preference is implemented as a tax break on domestic products used in government procurement, introducing a price difference between domestic and imported commodities of 10% to 20%, depending on the sector. To determine the level of the price difference to implement the scenario, sales taxes and tariffs are adjusted simultaneously, keeping import prices constant. The LCR under review applies to government procurement, so this modeling approach, influencing government income and expenditure, is a reasonable choice. The results depend on the quality of the data representing government consumption in the database, whereby statistics on government consumption are generally rather critical. To the extent that the government use is understated in the database, the results will be larger. In addition, given that the tax break is implemented on average across the various sectors, the results for any specific supply sector will vary. Results Local content in Brazil’s government consumption in the database is 5% for computers and electronic products, electrical equipment, and transport equipment Localization measures 37 and 26% for machinery. The policy doubles local content in these sectors. Local content is already high for motor vehicles (96%) and communication services (93%) and increases only marginally, to 97% and 95%, respectively. Government imports in the targeted sectors decrease between 5% and 25%. However, government consumption in the targeted sectors accounts for a maximum 0.01% of consumption across all use categories, so the government sector is too small in this area to have a visible effect on the aggregate or sector level. Thus, these results provide a good example where, while identifiable and transparent, LCRs may not have a notable difference on economic outcomes. Indonesia: LCR in the automotive industry The LCR is modeled as a 100% LCR on motor vehicles, represented by two sectors – “motor vehicles and parts” and “transport equipment” – for final demand. Imported motor vehicles being used as intermediates into production in the automotive industry are not covered by the LCR, assuming these are parts going into assembly in Indonesia. Thus, the measure is well depicted in the database and relatively straightforward to model. However, the model implies that all current operators comply with the measure. To the extent that operators leave the sector as a result of the policy, the results are understated. Results The automotive industry in Indonesia accounts for 4% of GDP, and 32% of automobiles for intermediate and final demand are imported (Table 2.2). The LCR increases the local content of motor vehicles from 63% to 89% and local content of transport equipment from 75% to 87%. As a result, imports of motor vehicles and transport equipment drop strongly across uses, by 68% and 44%, respectively (Table 2.6). Imports of intermediates of these commodities decrease in most sectors, except in the automotive industry, which increases imports of parts to satisfy domestic demand. As a result of decreasing automotive imports, the exchange rate appreciates to balance the current account, and imports in other manufacturing sectors increase between 1% and 4% (Table2.3). On the country level, imports and exports decrease by 2.3% and 2.5%, respectively, and the terms of trade improve by 0.7% (Table 2.2). The policy increases demand for domestic motor vehicles strongly, and domestic production of motor vehicle and transport equipment increases 42% and 23%, respectively (Table 2.6). Increasing production leads to increasing demand for labor in the automotive industry, and workers reallocate to the automotive industries and wages increase 0.2%. Returns to capital increase 0.1%, so that sectors that are labor and capital intensive experience increasing production costs. At the same time, import prices decrease. For inputs across other parts of the economy, import prices drop by 0.4% to 0.6% while, they decline 9% 38 Dorothee Flaig and Susan F. Stone Table 2.6 Indonesia – LCR in the Automotive Industry, Percentage Changes (i) Motor vehicles sector Quantity Prices Total Use category Use category mediate nment mediate nment Inter inputs Private consumption Gover consumption Capital goods Inter inputs Private consumption Gover consumption Capital goods Production 42% 110% 4% −4% 43% −1% −1% −1% −1% Exports 32% 57% 1% −5% 24% −6% −1% −1% −3% Domestic 45% 153% 5% 301% 47% 3% 0% 29% 0% Imports −68% −62% −56% −16% −91% −9% −8% −3% −20% Total demand 3% 7% 1% 0% 0% −1% 4% 0% (ii) Transport equipment sector Quantity Prices Total Use category Use category mediate nment mediate nment Inter inputs Private consumption Gover consumption Capital goods Inter inputs Private consumption Gover consumption Capital goods Production 23% 18% 4% −1% 238% 0% 0% 0% 0% Exports 28% 5% −3% −6% 109% −1% −1% −1% −5% Domestic 22% 19% 5% 301% 311% 0% 0% 18% 2% Imports −44% −3% −58% −16% −87% −1% −6% −2% −12% Total demand 6% 13% 0% 0% 0% 0% 2% 1% LCR = local content requirement. Source: Model results. for motor vehicle parts (transport equipment – 1%) due to a policy-induced decrease in import demand. This price decrease benefits sectors with large import shares of intermediates such as petroleum, chemicals, computers and electrical equipment, electricity, and water and air transportation. Motor vehicle parts are a major input, so the effect is especially strong for motor vehicles. Localization measures 39 Decreasing export competitiveness and increasing import competition, resulting from the currency appreciation, and increasing production costs in most other sectors, have a negative impact on production in non-targeted sectors (Table 2.3). Thus, on an aggregate level, production and GDP do not change. Saudi Arabia: local content and government procurement measure on products in medicines and medical supplies The measure is modeled as a 90% minimum requirement of domestic content for government procurement in the following sectors: textiles, wood products, paper products, chemicals, basic pharmaceuticals, rubber and plastic products, and electronic equipment. Allowing for products that are not subject to the measure, 10% of products in each sector are assumed exempt. As noted, accurate government procurement data are difficult to obtain. Thus, the model outcomes depend on the accuracy of Saudi Arabia’s government procurement in the METRO database. To the extent that it is understated, the impact on government services, and indeed the total impact, will be understated. Results Government procurement of the targeted commodities, of which 20% are imported, accounts for 0.7% of Saudi Arabia’s GDP (Table 2.2). On the commodity level, the LCR is binding for chemicals and pharmaceuticals, where local content increases from 66% and 40%, respectively, to 90%. Overall, government imports decrease 3.1%, while imports of all other uses increase slightly, leading to a total import decrease of 0.1% (Table 2.2). There is no noticeable effect on the exchange rate. Table 2.7 shows the effects exemplarily for pharmaceuticals. Imports are substituted by domestic products in government procurement and trigger domestic production. Other uses are only slightly affected and show contrasting tendencies, increasing imports, and lowering demand for domestic goods as the government increases prices for domestic products. Government consumption is large enough to increase production in the sectors where the LCR is binding – 0.4% for chemicals and 3.8% for pharmaceuticals. Non-targeted sectors experience a small but negative effect (Table 2.3). South Africa: new mining bill The measure is modeled as an LCR on inputs to mining, abstracting from complex company ownership requirements. The model captures these measures by depicting the policies with respect to the domestic content requirement of 70% on goods and 80% on services. Once again, we assume that all current operators comply with the measure. To the extent that companies 40 Dorothee Flaig and Susan F. Stone Table 2.7 Saudi Arabia LCR – Pharmaceutical Sector, Percentage Changes leave the market as a result of the policy, the impacts are understated. In addition, the extent to which the ownership structure impacts the degree of implementation will impact the results. Results Mining contributes 4% to South African GDP, and the sector imports 17% of its inputs (Table 2.2). The LCR is binding for most of the production inputs and increases domestic supply and production in sectors across the board, with the notable exceptions of metals and minerals (Table 2.3). The supply of mining products to the domestic market increases 2% (Table 2.8). While there is no noticeable effect on input prices or production costs in mining, its output declines by 0.8%. This stems from a decline in export markets. Mining is an important export sector in South Africa – 97% of its total output is exported, and it accounts for 12% of the country’s total exports. Exports decrease in response to relative exchange rate effects. As a consequence, mining exports decrease 0.9%, driving the decline in production. US: Buy America bill enacted by various states The US Buy America program is implemented through transportation grants across many US states. Thus, any public project funded by these grants must use some level of domestically produced iron, steel, and other manufactured goods. The amounts of both the grants and the LCR vary by state and by projects within states. Quantity Total Use category Intermediate inputs Private consumption Prices Use category Government consumption Capital goods Intermediate inputs Private consumption Government consumption Capital goods Production 3.8% −0.2% 0.0% 125.0% −0.1% 0.0% 0.0% 0.0% 0.0% Exports 0.5% −0.2% −0.1% 71.0% −0.2% 0.0% 0.0% −4.1% 0.0% Domestic 4.2% −0.1% 0.0% 127.1% −0.1% 0.0% 0.0% 0.2% 0.0% Imports −2.3% 0.1% 0.1% −83.4% 0.0% 0.0% 0.0% −12.5% 0.0% Total demand 0.0% 0.1% 0.0% 0.0% 0.0% −1.0% 0.0% LCR = local content requirement. Source: Model results. Localization measures 41 Table 2.8 South Africa LCR – Mining Sector, Percentage Changes Quantity Total Use Category Intermediate inputs Private consumption Government consumption Prices Use Category Capital goods Intermediate inputs Private consumption Government consumption Capital goods Production −0.8% −0.8% 0.1% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Exports −0.9% −0.9% 0.1% 0.1% 0.0% −0.1% 0.0% 0.0% 0.0% Domestic 2.4% 2.4% 0.0% 0.1% 0.0% 1.8% 0.0% 0.0% 0.0% Imports −0.8% −0.8% 0.0% 0.0% 0.0% −0.2% 0.0% 0.0% 0.0% Total demand −0.1% −0.1% 0.0% 0.0% 0.0% 0.0% 0.0% LCR = local content requirement. Source: Model results. The model does not provide for measurement of individual grants at the subnational level. However, looking across hundreds of grants, accounting for the size of the grant and relative size of the transport budgets, this measure is modeled as an average increase in domestic content of 25 percentage points in total government spending on transport equipment and services. Thus, domestic content is raised from 61% in motor vehicles, 5% in transport equipment, and 55% in transport services, to 86%, 30%, and 80%, respectively. This is an average across hundreds of individual policies and grants and thus represents an average impact. Results Imports account for 49% of US government procurement spending in transport equipment and services (Table 2.2). The measure increases domestic demand in the targeted sectors and reduces imports. However, government procurement in this area accounts for only 0.002% of GDP, so the effect is not large enough to visibly impact on the sector level (Table 2.9 presents the example of motor vehicles). Again, while the measure is transparent and identifiable, it illustrates the case where large domestic markets can often afford to implement this policy with small measurable side effects. However, what the model fails to capture in this instance is the longer-term impacts this policy has on innovative behavior by firms. By having access to government contracts on a noncompetitive basis, firms have little incentive to invest in innovative or cost-cutting behavior. This reduces their competitive stance vis-à-vis firms that face contested markets. DOI: 10.4324/9781003415794-3 This chapter has been made available under a CC-BY-NC-ND 4.0 license. Objectives of the chapter Historically, the mining industry in resource-rich developing nations has operated as an enclave, extracting raw materials for export with few links with other sectors and little value added to the resource-rich country. This diminished the opportunity for direct economic and social benefits, causing many mining countries to endure undiversified economic structures, high unemployment rates, and macroeconomic frameworks vulnerable to commodity shocks. Legal and fiscal frameworks adopted during the 1990s focused on the capacity of the mining sector to generate tax and royalty revenues whose benefits, it was assumed, would automatically trickle down to the rest of the economy (Bastida, 2014). However, this approach failed to appreciate that procurement of goods and services is the single largest in-country economic expenditure over the life of a mining project – sometimes larger than taxes, salaries, wages, and community investment combined.1 At present, global mining companies follow high procurement standards and tend to outsource their operational activities to globally competitive contractors. The move toward greater outsourcing has led to the emergence of global supply chains in the mining sector, enabled by falling transportation costs, lower trade barriers, improved information and communication technologies, and liberalized financial regulations (Östensson, 2017). Since the procurement of goods and services is the single largest expense over the life of a mining project, local content requirements (LCRs) have become politically attractive, as they aim to respond to demands to create jobs and economic opportunities. Yet in practice, the design and implementation of policy frameworks for LCRs in mining that foster sustainable and competitive domestic suppliers has proven extremely difficult. Many countries have introduced or amplified targets for locally supplied goods and services in legislative and regulatory instruments without surveying mining companies’ procurement needs, establishing a baseline of local supply capabilities, or calculating the trade-offs from specific initiatives in terms of value created for the economy. Mining companies argue that such targets are often prescribed without due consideration of the complex operational 3 Local content policies in the mining sector Jane Korinek and Paulo De Sa Local content policies in the mining sector 49 structures, strategies, and market conditions of the industry, and they have often responded with caution or sought to circumvent compliance with such targets. Notwithstanding, many companies have set up internal local content programs that also serve as risk-mitigation actions and have developed constructive engagement with governments and communities to reinforce their social license to operate. This chapter draws on some observations regarding the effectiveness of LCRs’ contribution to foster employment and economic diversification and increase government revenues in resource-rich countries. It addresses policy implications for countries that aim to maximize benefits from the mining sector while ensuring that their regulatory and business environments contribute to sustaining the sector’s competitiveness in global markets. The chapter begins with an introductory section that presents current definitions for LCRs and suggests a typology of local content policies that are used in the mining sector and some considerations when aiming to measure their impacts. The second section reviews local sourcing and domestic employment requirements, trade restrictions, and other local content measures that have been introduced in different countries to foster backward and forward linkages and analyses common pitfalls in their design and implementation. The third section includes reflections on institutional frameworks relevant for countries that aim to deploy LCRs. It includes a brief discussion on the implications of World Trade Organization (WTO) rules and investment agreements on the use of such instruments. The fourth section relies on a qualitative desk review of existing literature, research papers, and reports on LCR law and practice that aims to shed light on some of the potential economic impacts of the use of LCRs in resource-rich countries. It highlights difficulties in measuring the impacts of LCRs, due in part to a lack of data providing empirical evidence but also because the definition of LCRs varies from country to country, making comparisons difficult. Notwithstanding, empirical evidence suggests that LCRs seldom achieve their objectives of increasing the domestic value added in the supply of goods and services or developing sustainable linkages but have been somewhat successful in some cases in terms of job creation and skills transfer. The last section presents a series of policy implications derived from the preceding analysis. It finds that LCRs, if implemented, must be aligned with what can be realistically achieved without threatening the long-term competitiveness of the industry. Therefore, they should be part of a broader set of public policies to leverage the sector’s contribution to increase value addition throughout the economy, job creation, and economic diversification rather than the share of domestic procurement. Accordingly, many resource-rich countries have moved from prescribed local procurement requirements to creating the conditions to increase the export of mining-related services and integration in global supply chains. 50 Jane Korinek and Paulo De Sa Scope and content of LCRs LCRs in the mining sector include all laws and regulations that prescribe measures to stimulate the use of locally sourced goods and services, create job opportunities, and generate broader spillover effects in the national and local economies of resource-rich countries. Their scope ranges from mandatory employment targets to tax exemptions on local procurement but also includes export restrictions to encourage downstream processing of domestic minerals, ownership requirements, the reservation of certain procurement from domestic firms, and demands that research and development (R&D) on mining-related technologies be conducted in the country where operations take place. An estimated 90% of resource-rich countries employ LCR, most of which are quantitative targets or requirements (McKinsey, 2013). There is no commonly agreed definition of what constitutes “local content.” The term “local” has been used in different ways (Korinek and Ramdoo, 2017; Intergovernmental Forum on Mining, Minerals, Metals and Sustainable Development [IGF], 2018c): it can refer to geographic proximity, such as the population living in the vicinity of a mining project, although local suppliers are often defined as businesses registered in the country at both the national and community levels. Similarly, local employment is generally associated with the nationality of the workforce and can target a wide range of job positions, including at the management level. In some countries, goods and services must be provided by firms with some percentage of domestic capital ownership. In others, mining companies may be requested to enter into partnerships with state-owned entities or local firms or to list a prescribed percentage of their shares on national stock exchanges. The definition and scope of what constitutes “content” can vary from country to country but usually aims to accomplish one or more of the following: (i) upstream supplier development at the domestic and/or community level; (ii) skills enhancement at different stages of the value chain, including the creation of job opportunities and training of the local workforce; (iii) technological and knowledge transfer, in particular toward local small and medium-sized enterprises (SMEs) or government agencies, or investing in R&D activities in the country; (iv) shared ownership of assets; and/or (v) downstream value addition and beneficiation of locally produced raw materials (Korinek and Ramdoo, 2017). Typology of main LCRs affecting the mining sector LCRs are generally formulated in policy frameworks setting broad orientations– such as national development plans and policy statements – and are codified in legislative and regulatory instruments or as part of contract agreements negotiated with mining investors. There is a wide range of policies to promote local content, but in simple terms, they can be classified as either demandside or supply-side policies. Demand-side policies focus on creating demand Local content policies in the mining sector 51 for locally procured goods and services. Supply-side policies emphasize skills development and building the capacity of local suppliers to bring them up to global standards of competitiveness on price, quality, and reliability. Measures range from mandatory or voluntary supplier development programs to standalone or public–private partnerships for skills development. Measures can be further classified as mandatory (“requirement-based” approach) or rely on firms’ voluntary programs or “best efforts” to grant preferences to local economic entities (“incentives-driven” approach). “Requirement-based” policies can be further classified into two categories. Some impose legally binding targets on firms, either in terms of quantity (e.g., the number of local staff to be employed or contracts to be awarded to local suppliers) or value (e.g., a percentage of total spending on local procurement). “Incentivesdriven” requirements generally do not set specific targets but can also be binding. For example, companies may be requested to publish their procurement needs or report on the percentage of local employment, although the levels of local procurement and employment might not be prescribed. “Best efforts” provisions are usually embedded in legislation but do not subject the companies to any specific quantitative requirements. LCRs can be further classified as aiming to foster greater linkages upstream of the mining sector, or downstream (Table 3.1). They contain measures aimed at developing job skills and the capacity of local suppliers, using the sector as the anchor client in the case of backward or upstream linkages or as the source of inputs in the case of forward or downstream linkages. In the case of LCRs that aim to foster backward linkages or increase local procurement, local sourcing and domestic employment requirements are the most common examples of mandatory LCRs. The first often mandates the purchase of specific product categories or a prescribed volume or value of goods and services from local suppliers. They can also include tender preferences for local suppliers and are increasingly accompanied by the obligation to provide procurement plans, schedules, and implementation reports to local authorities (World Bank and Kaiser Economic Development Partners, 2015). Domestic employment requirements call for the hiring of specific percentages of local workers and for companies to reserve some categories of jobs exclusively for nationals (and, increasingly, for indigenous people, women, or disadvantaged groups). They can also limit the number of expatriates employed and mandate training programs for their replacement by local workers. Governments can act as facilitators, providing incentives to firms to increase their local purchases. Supplier development programs (SDPs) are by far the most common government “incentives-driven” approach to increase upstream or backward linkages; other approaches include (i) tax reductions for firms that are part of mining supply chains, support workforce development, invest in innovation, or agree to transfer technology; (ii) grants and scholarships for students and employees willing to upgrade their skills or for training institutions that partner with industry to develop them; and (iii) support for 52 Jane Korinek and Paulo De Sa Table 3.1 Illustrative Examples of LCRs Type Demand Side Supply Side Policies that aim to increase upstream linkages Requirement Extractive firms are Extractive firms are required to: based required to: • Provide training to employees • Meet numerical • Fund capacity development programs targets for local • Establish a “buddy system” whereby employment per type local staff are paired with expatriates of jobs and/or level for direct on-the-job training of competency or report on measures taken to hire locally • Extend preferences to local suppliers for procurement of goods and services Incentives Extractive firms Extractive firms are required or driven are required or encouraged to: encouraged to: • Conduct training programs for • Publish job vacancies potential suppliers to understand • Publish tenders on the needs and required standards of given websites and in extractive firms the media • Conduct awareness campaigns • Set up or use existing about key procurement employment networks of suppliers opportunities Governments and academic institutions: • Create engineering and technical curricula in conjunction with extractive firms’ stated and future needs • Provide targeted training to enhance the skills of potential suppliers Governments or regulators: • Set up networks of suppliers and extractive firms • Provide forums for matchmaking between local suppliers and extractive firms to foster greater engagement Policies that aim to increase downstream linkages Requirement Extractive firms are Extractive firms are required to: based required to: • Invest in downstream processing • Sell a share of facilities their raw materials Concomitantly, governments or in-country regulators may: • Pay higher tax • Provide tax concessions, concessional rates on exports of loans, and lower import duties raw materials than on imported capital goods or processed products subsidize energy, transport, or other • Engage in infrastructure downstream processing in order to obtain export licenses Local content policies in the mining sector 53 Type Demand Side Supply Side Incentives driven • Extractive firms are given tax concessions if they favor incountry sale of raw materials • Extractive firms are required to divest a share of their equity if they do not process raw materials in-country Extractive firms are required or encouraged to: • Collaborate with training institutions to ensure the necessary skills for processing industries Governments or regulators: • Provide tax or other concessions to processors to invest LCR = local content requirement. Source: Authors’ conception; Korinek and Ramdoo (2017). research, development, and innovation through dedicated funding to universities, research centers, and innovation incubators. Notwithstanding, several countries have introduced prescriptive beneficiation requirements either in the form of across-the-board legislation or during negotiations of contracts with mining firms. These requirements are essentially of three types: (i) taxes on the export of unprocessed minerals; (ii) quantitative export restrictions on unprocessed minerals, either partial (quotas) or comprehensive bans; and (iii) export licensing requirements to control mineral exports (UNCTAD, 2017; Fung and Korinek, 2013). In other countries, governments have created or mandated existing state-owned enterprises to invest in the downstream sector. China is the most obvious example, with state entities owning and operating most copper processing facilities and supporting them with direct grants, low-interest loans, and tax incentives (Geipel, de Weerdt, and Alarcon, 2021). Prescriptive beneficiation requirements are often accompanied by incentives to increase the rate of return of downstream investments such as tax reductions or exemptions, energy or water subsidies, and concessional loans or the provision of industry-specific infrastructure such as industrial parks. Governments may also offer protections for processing operations through customs tariffs and import restrictions on the products being processed (IGF, 2018c). Measuring the impacts of LCRs The economy-wide costs generated by local content measures should be measured and compared with the potential benefits they aim to provide. LCRs on intermediate inputs to the extractive industries may lead to an increase in production costs that will raise output prices, in the mining sector in the first instance.2 The increased prices of raw materials raise costs 54 Jane Korinek and Paulo De Sa for producers further down the value chain, reducing the competitiveness of downstream industries and ultimately hindering the development of the wider economy (Grossman, 1981) and economic diversification. The size of these efficiency losses will be proportional to the additional costs associated with purchasing required inputs domestically, due to the policy, compared with their cost on international markets (Stone, Messent and Flaig, 2015). In addition, such measures can reduce the amount of taxes collected by the government if they negatively affect the profitability of firms and hence shrink the tax base. Using mandatory quantitative requirements to develop internationally competitive industries runs counter to the known positive spillovers from engaging in global value chains and the role trade plays in their development. Kimura and Obashi (2011), for example, argue that the success of global value chains in East Asia, especially compared with Latin America, relies heavily on such interlinkages between domestic and foreign markets. Impacts on the investment climate of mandatory LCRs should also be measured. One of the central tenets to an attractive investment climate is nondiscrimination (OECD, 2015). Restrictions on foreign direct investment (FDI) and trade have been found to result in less FDI overall. LCRs: some examples Prescribed demand-side requirements that aim to create backward linkages have been used in many African countries recently; the cases of Ghana, South Africa, Tanzania, and Zambia are outlined here. Countries such as Australia, Canada, and Chile have preferred an incentives-based approach to develop competitive suppliers and strong linkages between mining and the domestic economy. Prescriptive requirements that aim to foster downstream minerals processing have been used in many countries; the cases of Indonesia and Botswana are discussed here at some length. Demand-side requirement-based LCRs that aim to create backwardlinkages First-generation local content legislation usually required “best efforts” from mining companies. More recently, especially after 2018, a growing number of African countries have adopted or revised mandatory LCRs in their mining laws – introducing mandated quantitative procurement targets, sometimes requiring foreign investors to open equity to local partners. This is the case of the Democratic Republic of Congo, Ghana, Namibia, South Africa, and Tanzania. Countries like Burkina Faso, Mali, Kenya, Mozambique, Nigeria, and Zambia are revising their mining and investment codes with the same purpose. The efforts to promote demand-side, requirement-based LCRs in Africa through mining legislation provide valuable lessons. Firstly, they have created Local content policies in the mining sector 55 a complex landscape across the continent that is difficult to navigate. Moreover, many countries have prescribed levels of local procurement beyond what local suppliers are capable of meeting, and few efforts can be considered completely successful. Many of the quantitative targets are aspirational and have little chance of being successfully implemented without mechanisms to build the capacity of current and potential local suppliers (African Natural Resources Centre, 2021). The following cases illustrate the challenges of designing and implementing LCRs in countries where, on many occasions, laws and regulations are frequently revised and retracted, creating complex regulatory frameworks and institutional ambiguity. (i) The case of Ghana Ghana’s 2006 Minerals and Mining Law sought to facilitate production linkages, and the 2012 Minerals and Mining (General) Regulations set quotas and timelines for compliance applicable to both mining companies and suppliers regarding employment of the local workforce and procurement of locally produced goods and services. The regulations had five main features (IGF, 2018a): • Numerical employment targets for the number of allowed expatriates, with restrictions for certain categories of positions reserved for local staff, and timeframes for compliance. Ghana’s regulations require 100% local workers for administrative or labor positions. For more technical, specialist, or management roles, restrictions on how many foreign nationals can be employed are set on a case-by-case basis. • Mandatory procurement of locally produced goods and services specified by a list published by the Minerals Commission. A first list, published in 2014, included eight products required to be sourced locally. In 2016, the number of products increased to 19, and currently 28 products are featured. • Compulsory reporting requirements mandating mining companies to submit a five-year local procurement plan stating how much they will buy from local firms and report on it each year, including progress on the listed goods and services. • Use of a phased approach with lead times to allow industrial capacity to meet the industry requirements for product cost, quality, and quantity. • Sanctions for noncompliance with the LCRs, which is assessed annually. The latest assessment indicates that local companies have managed to supply, on average, about half of the products specified in the Minerals Commission list, increasing local procurement from $148 million to $394 million between 2014 and 2018 (Atta-Quayson, 2022). However, these products are not necessarily produced locally since mining companies can meet LCRs for listed 56 Jane Korinek and Paulo De Sa goods by purchasing from resellers. Regarding local employment, most companies have been able to meet the quotas set in the legislation for all professional categories listed (IGF, 2018d).3 (ii) The case of South Africa South Africa has one of the world’s most complex local content legislations. The 2000 Preferential Procurement Policy Framework enables the designation of specific sectors for preferential domestic manufacture (or “localization”), in line with national development and industrial policy goals. The 2011 Beneficiation Strategy for the Minerals Industry identified a range of crosscutting constraints to local beneficiation and proposed a series of policy, legal, and regulatory measures to increase value added in the mining sector and facilitate job creation, industrialization, and economic diversification. The Broad-Based Socio-Economic Empowerment Charter for the Mining Industry (amended in 2010), commonly known as the Mining Charter, aimed to rectify the results of discrimination based on race, sex, and disability. It defined “Black Economically Empowered” (BEE) entities as those where historically disadvantaged persons hold a minimum of 25% plus one vote of capital. The charter requires the industry to procure from BEE entities according to set targets and to track procurement, employment, and other criteria in line with social transformation targets (de Weerdt and Geipel, 2020). The Chamber of Mines established a Broad-Based Scorecard, which requires its members to comply with a procurement target from BEE entities of 70% and 80% for goods and services, respectively. Targets were then broken down into further subcategories with certain percentages devoted to different groups, including historically disadvantaged persons, BEE-compliant companies, and womenor youth-owned companies. Mining companies were given an implementation period of 5 years for goods, with interim targets leading to full implementation (Geipel, de Weerdt, and Alarcon, 2021). While many of the 2010 charter’s targets appear to have been met, they were essentially based on company ownership rather than the level of value addition generated by a specific business. Moreover, there was evidence that locally produced products for the mining sector at times were displaced by imports sold by businesses compliant with the targets (Korinek and Ramdoo, 2017). As a result, the 2018 Mining Charter stipulated that 70% of all goods used in the mining industry must be manufactured in South Africa, within 5 years of the law’s enactment, and not merely purchased from South African–registered suppliers. Yet mining companies have argued that the target was set too high and that there are not enough supply-side policies to support them. Overall, the disconnect between different definitions across various pieces of legislation, frameworks, and scorecards – and the lack of open dialogue and trust between government and industry – increased the uncertainty around Local content policies in the mining sector 57 the interpretation of the targets and the overall success of the policy (Moraka and van Rensburg, 2015; White, 2017). It was determined that some of the complex ownership and participation requirements were not implemented in good faith, and a Code of Good Practice for the South African Mineral Industry was devised to set out administrative principles for the effective implementation of the mining legislation and the Broad-Based Socio-Economic Charter applicable to the mining industry. The code defines ethics of conduct to ensure the Mining Charter is implemented in good faith and to prevent abuses such as fronting practices and opportunistic behaviors that may divert potential benefits from the targeted stakeholders. For example, the Code of Good Practice defines practices considered fraudulent, such as situations in which local stakeholders may be appointed to a position but discouraged or inhibited from participating in core activities; economic diversion, in which economic benefits received do not flow back to the local stakeholder in the ratio specified in the legal document; and intermediaries leveraging their BEE status that have concluded agreements with mining companies (fronting operations) (Korinek and Ramdoo, 2017). (iii) The cases of Tanzania and Zambia In Tanzania, the 1997 Mineral Policy, the 2010 Mining Act and 2017 amendments – Written Laws (Miscellaneous Amendments) Act – emphasized the development of backward linkages but left implementation largely to voluntary compliance. In January 2018, the Minister for Minerals promulgated the Mining (Local Content) Regulations, introducing hard quotas for the procurement of goods and services from domestically owned providers, including in the banking, financial services, insurance, and legal sectors. Mining companies are now required to prepare an annual local content plan, including projections for procurement, employment, and training activities. The plan must be updated annually and submitted for approval to the Local Content Committee. It is also a requirement to prepare a revolving 3to 5-year program for R&D, detailing planned activity expenditures and calls for proposals for their implementation. Licensees also need to publish a plan for technology transfer to the benefit of Tanzanian entities. The regulations also stipulate that mining companies can only retain the services of Tanzanian financial entities and need approval of the Mining Commission to hire the services of foreign financial institutions. They must maintain a bank account and conduct business with a bank with majority Tanzanian shareholding and may only retain the services of Tanzanian legal practitioners whose principal office is in Tanzania (Herbert Smith Freehills, 2018). Difficulties in implementation led the government to pass in February 2019 the Mining (Local Content) (Amendment) Regulations of 2019, which amended the 2018 Regulations in a number of ways. Among other things, the amended regulations reduced ownership restrictions for financial institutions preference 64 Jane Korinek and Paulo De Sa growth in the supply of refined nickel for electric vehicle (EV) batteries over the next decade.15 The nickel export ban increased downstream processing because Indonesia’s exceptional resource endowment provided strong incentives for investments, mostly from Chinese companies (Lebdioui and Bilek, 2021). However, its standing among Western mining investors has fallen sharply: Indonesia is ranked the fourth worst mining jurisdiction globally (out of 78 jurisdictions), outranked only by Venezuela, the Chubut province of Argentina, and Tanzania, according to a survey of mining investors (Yunis and Aliakbari, 2020). The Indonesian government is preparing a new nickel strategy, considering the possibility of levying an export tax on products with less than 70% nickel content, and limiting the construction of smelters for class 2 (lower-grade) products. This strategy is part of Indonesia’s policy approach for the energy transition (Huber, 2021). The country aims to develop a fully integrated domestic supply chain for nickel, from ore extraction to battery production and EV assembling. The government has produced a road map for EV battery development and storage systems through 2026. For example, in 2021, LG Energy Solution and Hyundai from the Republic of Korea (henceforth, Korea) jointly started building a $1.1 billion EV battery plant in West Java. In addition, the Chinese company Huayou is involved in several smelting projects on Sulawesi Island, including two projects with the Indonesian unit of Vale, estimated to cost around $6.3 billion. Ford Motor Company is negotiating its involvement in one of these projects. Huayou is also teaming up with Tsingshan and Volkswagen Group China, with the goal of supplying nickel and cobalt from Indonesia to support the production of batteries.16 Battery manufacturers are also investing. In April 2022, China’s largest battery maker, Contemporary Amperex Technology Co., Limited (CATL), PT Aneka Tambang, and PT Industri Baterai Indonesia (IBC) signed an agreement to develop a project in Indonesia’s North Maluku Province that will focus on nickel mining and processing, battery materials, and battery manufacturing, as well as battery recycling. CATL is investing in Indonesia through QMB New Energy Materials, a joint venture with Tsingshan and Chinese battery recycler GEM. Korean battery maker LG Energy Solution is separately partnering with IBC and Aneka Tambang to develop an end-to-end battery supply chain in Indonesia. When these new foreign investments materialize, Indonesia will account for around half of the world’s growth in nickel production between 2021 and 2025, according to the International Energy Agency, and could become a leading producer of nickel-based products, including EV batteries. The expectation in Indonesia is that large investments will increase the economies of scale and drive down costs, potentially making Indonesia a low-cost manufacturer, competitive in global markets. Local content policies in the mining sector 65 (ii) The case of Botswana Diamonds were first discovered in Botswana in 1966, shortly after independence, with large-scale production starting in 1971. Diamond mining became the greatest contributor to gross domestic product (GDP) (currently around 30%) and government tax revenues (currently around 60%) (Columbia Center on Sustainable Investment, 2016). In the early 1980s, the government of Botswana tried to promote the development of a diamond-cutting and -polishing industry. However, global mining company De Beers, which dominated production in Botswana and the sale and marketing of diamonds globally, did not support this ambition, arguing that cutting and polishing activities were not economically viable in Botswana. Three cutting and polishing factories were established between 1980 and 1990, but none of them ever reported a profit. A second opportunity emerged in 2005, when De Beers’ 25-year mining license was due for renewal. Botswana’s negotiating leverage derived from De Beers’ reliance on Debswana, a 50–50 joint venture with the government, which was responsible for about 60% of the company’s global supply of rough diamonds. Botswana obtained a guarantee from De Beers that a percentage of the diamonds mined in the country would be allocated to national cutting and polishing companies and that all sorting and valuing operations would be undertaken in-country (UNIDO, 2012; Korinek, 2014). Subsequently, the government invited foreign cutting and polishing companies to set up operations in the country with the promise of a guaranteed long-term allocation of De Beers’ diamonds at 20% to 30% below the market price, on the condition that they hire and train locals with cutting and polishing skills (IGF, 2018c; 2018d). Furthermore, a Diamond Academy was opened by the joint venture to train diamond sorters and valuation staff (Korinek, 2014). In 2008, the Diamond Trading Company (DTC Botswana), a 50–50 joint venture between De Beers and the government, was established to sort and value Debswana’s output and manage the supply of diamonds to the domestic cutting and polishing industry. De Beers also agreed to move its aggregation business – selecting and mixing the diamonds from De Beers mines for its customers – from London to Gaborone, in the hope of creating considerable spillovers to other industries such as hospitality, finance, and transportation, since diamond buyers would now have to go to Gaborone to buy De Beers’ diamonds (Morris, Kaplinsky, and Kaplan, 2011). Nonetheless, a remaining issue is the ability of Botswana’s diamond-cutting and -polishing industry, which remains dependent on government incentives, to compete with low-cost factories in Asia, especially in India. Unless investments in support infrastructure decrease production costs substantially, Botswana’s diamond-polishing sector may not survive in the longer term. Moreover, its diamond processing industry is built on access to raw diamonds, but domestic diamond reserves are expected to be exhausted in 30 to 40 years (Columbia Center on Sustainable Investment, 2016). 66 Jane Korinek and Paulo De Sa Institutional frameworks and coordination High-quality institutions in charge of designing, governing, managing, enforcing, monitoring, and evaluating LCRs are vital to accomplish their objectives. Many regulatory entities do not have a deep knowledge of the mining supply chain and its potential to generate revenues, businesses opportunities, and employment. Many are understaffed or underfinanced and suffer from opaque decision-making processes. The effectiveness of LCRs can be affected by gaps in the institutional framework and lack of coordination for the approval of cross-sector policies and regulations, responsibility for which is scattered across many ministries, which often operate in silos. Therefore, some countries have established dedicated local content entities to coordinate and monitor progress, drawing on the resources of all relevant government agencies. This is the case of the Local Content Committee in Tanzania and the Industry Participation National Framework Authority in Australia. Reporting requirements are important tools to monitor LCRs, but their efficacy depends on the capacity of the entity in charge of enforcing them. Because many countries have yet to enact and enforce adequate reporting mechanisms, in practice, most existing LCRs enable noncompliance (White, 2017), as shown in some of the previous examples. To address this, governments are increasingly asking mining companies to submit local procurement plans as part of their yearly reporting requirements. Yet sometimes, these reports cover only the firms’ local capacity-building activities and are not suitable to monitor progress toward achieving broader desired outcomes. Insufficient collaboration between governments, the business community, and civil society can lead to situations in which procurement targets are set at unrealistic levels, causing enforcement difficulties or creating conditions that facilitate influence peddling or corruption of public officials. LCRs can provide preferential treatment to stakeholders with strong vested interests or that are politically affiliated and may engage in rent-seeking behavior; some examples of this effect are outlined above. Likewise, fear of competition may block cooperation among mining companies, leading to a duplication of uncoordinated initiatives that prevent local suppliers from achieving the scale needed to become competitive. Partnerships between mining firms, governments, training institutions, and local stakeholders are important, in particular when implementing SDPs, as they tend to be more effective when they benefit from the participation of multiple institutional actors, including notably mining firms (African Natural Resources Centre, 2021). For example, some subnational governments offer technical assistance to SMEs in the procurement contract process or keep databases of local suppliers to reduce information gaps that diminish their chances of responding to tenders, thereby combining supply-side and demand-side policies. Local content policies in the mining sector 67 WTO rules, investment agreements, and LCRs LCRs and other measures to increase the use of domestic goods and services, such as trade restrictions, subsidies, tariffs, and tax incentives, can introduce distortions in favor of local producers and may therefore contravene a number of trade and investment agreement disciplines (see Annex Tables 3.A1 and 3.A2). The relevant WTO commitments that relate to LCRs (Columbia Center on Sustainable Investment, 2016; Korinek and Ramdoo, 2017; Korinek and Bartos, 2012) are: • The National Treatment Obligation (Article III of the GATT) clause prevents governments from discriminating between like products from local industries and imports. This applies to policies such as those that force foreign companies to buy goods or services produced by locally owned companies or hire local service suppliers. • The Agreement on Trade-Related Investment Measures (TRIMs) prohibits the use of most forms of performance requirements on goods, set out in an “illustrative list.” These apply to domestic sourcing requirements, either in the form of lists of goods or quotas or percentages, as well as requirements to sell products domestically. However, developing countries have derogations to some of those commitments outlined in Article XVIII of the GATT. • Article XI:1 of the GATT imposes a general ban on quantitative export restrictions, but Article XI:2 and Article XX offer broad-scope exemptions to the ban on export quotas. Notably, Article XX(g) allows for quantitative restrictions relating to the conservation of exhaustible natural resources on the condition that “such measures are made effective in conjunction with restrictions on domestic production or consumption.” • The SCM Agreement prevents governments from providing incentives and granting subsidies that are contingent on sourcing goods domestically. Common subsidies, such as targeted tax preferences, may also be actionable under WTO rules. • The General Agreement on Trade in Services (GATS) regulates LCRs with regard to foreign investment and employment. Local equity requirements and employment quotas are generally prohibited, but these are only regulated to the extent that countries have taken specific commitments. Despite clear rules prohibiting certain forms of LCRs, many countries maintain them or have introduced new ones in recent years. Countries that joined the WTO after 1995 have been subject to closer scrutiny and tighter obligations, and some of them, such as Kazakhstan, have agreed to remove many LCRs during their accession negotiations (Korinek and Ramdoo, 2017). Measures imposing LCRs have been the subject of significant exemptions for developing countries within WTO rules, including to support infant industries and address balance-of-payments problems. Moreover, TRIMs apply to 68 Jane Korinek and Paulo De Sa restrictions on goods; regarding services, they only apply to commitments contained in countries’ GATS schedules. Few complaints regarding LCRs in the extractive industries have been brought to the WTO, although some have been filed regarding export restrictions (see claims regarding Indonesia and the case against China described earlier). Disputes regarding the enforcement of LCRs go through the WTO’s settlement system, which is usually a costly and long process (African Natural Resources Centre, 2021); moreover, the WTO dispute settlement system has been effectively halted in recent periods. International and bilateral investment treaties often go beyond WTO restrictions, e.g., by including restrictions on performance requirements for technology-transfer and R&D programs to be conducted in-country. They can also contain fair and equitable treatment obligations preventing governments from interfering with foreign investors’ “expectations” for their operations. Most treaties do not include these provisions but, when included, the obligations tend to be implemented more frequently, since these agreements usually use investor-state arbitration rather than the WTO’s state-to-state arbitration system, which increases the likelihood of complaints being filed (Columbia Center on Sustainable Investment, 2016). Some reflections regarding the economic impacts of LCRs Many LCRs lack a broader policy framework, such as overall political economy objectives, policy statements, and national development plans, to support them. Many countries have adopted stringent targets in regulatory instruments that lack detail and clarity and have a narrow scope of objectives. Before introducing them, governments must ascertain where the mining sector fits in relation to national development objectives, including its potential contribution to foster employment, government revenues, and economic diversification. Demand-side, requirement-based LCRs often aim for high percentages of local content without developing a detailed view of procurement spending, establishing a baseline of local suppliers’ capabilities, and quantifying the trade-offs from specific initiatives in terms of value added created (Elborai et al., 2019). Inaccurate understanding of mining firms’ procurement needs and of the absorptive capacity of local suppliers have led to the prescription of unrealistic targets, set beyond levels that local firms can meet, especially in countries with a weak industrial base and a private sector that is small, informal, and with low productivity. This can be a deterrent to the mining industry, especially without supporting government-sponsored supply-side measures. This was one of the issues confronted by Ghana’s many LCRs applied to the mining sector. While a broad definition of local content provides more flexibility for firms to meet targets and objectives, it is difficult to assess to what extent LCRs create local value-added and spillover effects for the rest of the economy. Ownership Local content policies in the mining sector 69 requirements do not always yield the best outcome in terms of domestic value added (ACET, 2017). Local ownership is not a relevant factor as long as companies create economic opportunities, and employment, and improve local labor skills (Esteves et al., 2014). Foreign-owned but locally based businesses can add value to the local economy, whereas local sourcing of imported goods usually does not. Duty-free imports of inputs by domestically owned firms place potential local producers and suppliers of such inputs at a disadvantage. For example, in less developed countries with a weak industrial base, many mining firms may report a relatively high percentage of local sourcing while, in reality, a large proportion of purchases originate from imports by local firms or representatives of foreign suppliers. This was found in Kazakhstan before it removed many of its LCRs; correspondingly, when the LCRs were removed, few local jobs were lost or displaced. Ownership requirements can also create an environment conducive to lack of transparency, corruption, and favoritism, where benefits may be captured by local elites embracing rent-seeking behaviors and failing to encourage entrepreneurial development. They tend to be ineffective, as foreign companies may bypass limits imposed on dividend distribution and exert effective control of joint ventures through shareholders’ agreements. Experience with such practices in South Africa prompted the authorities to institute a Code of Good Practice for the South African Mineral Industry to define ethics of conduct and prevent abuses such as fronting practices (Korinek and Ramdoo, 2017). Prescriptive beneficiation (downstream processing) requirements have often failed because there is little guarantee that domestic processing industries can become competitive in the longer term. For example, in late 2019, Zambia was forced to end a 15% export tax on raw gemstones because the tax was decreasing investments in the sector, and overall production had fallen. Another high-profile case was Tanzania’s ban on exports of raw gold, silver, copper, and other metallic minerals starting in March 2017. The ban did not achieve its goal of having mining companies invest in processing facilities, and raw mined gold stockpiled in the country until the ban was lifted (Geipel, de Weerdt, and Alarcon, 2021). Another illustration among many is the Indonesian bauxite sector, which has not recovered its exports since the export ban on bauxite was instituted in 2014. The examples of Indonesian nickel and Botswanan diamonds show, however, that governments have more leverage in negotiations with foreign investors regarding prescriptive beneficiation requirements when the country possesses an exceptional resource endowment. In the case of Indonesian nickel, the draw of a substantial share of global reserves of high-quality ore, high projected demand for the mineral for batteries, and continued dependence on access to the mineral by China forced investments in downstream smelters. In Botswana, the favorable conditions guaranteeing access to cheaper rough diamonds granted to downstream cutting and processing firms incentivized investors to open facilities. However, even these examples call into 70 Jane Korinek and Paulo De Sa question the longer-term sustainability of the downstream processing operations – in Indonesia for reasons of investment climate and in Botswana due to falling diamond reserves and lower productivity among diamond-cutting and -polishing operations compared with international competitors. Moreover, the global market for the specific mineral must present very favorable trends, with high projected future demand threatening a potential situation of scarcity, for such policies to be feasible even in the medium term. Governments are more likely to achieve their aims if they are bargaining from a favorable position either because they have strong negotiating capacity or because the mining company needs to renew the terms of its license or concession in a mine that is an essential component of its profitability. Botswana has expanded its reach through the value chain of diamond valuation, aggregation, and sorting in a step-by-step fashion, in the context of subsequent license and permit negotiations (Korinek, 2014). This gradual approach helped to ensure the industry partner was supportive of the policies and that responses were found to supply-side constraints, such as the opening of a Diamond Academy to train potential employees for diamond sorting, valuing, and aggregation jobs. Although empirical evidence suggests that “best efforts” clauses do not have much impact on their own (Geipel, de Weerdt, and Alarcon, 2021), the reality on the ground suggests that mandatory quantitative LCRs have yet to generate significant amounts of locally sourced inputs or strengthened intersector linkages. Local employment and some training of the local workforce appear to have been their most successful outcomes (Ellis and McMillan, 2020; White, 2017). Supply-side initiatives appear to work better than prescriptive targets when they focus on building workforce skills – through both targeted training and skills-transfer activities, including in business management skills – and the capacity of local businesses to supply goods and services competitively and integrate in global supply chains (McCulloch et al., 2017). Some of the most interesting examples come from public–private partnerships that combine mining firms’ SDPs with government-sponsored skills transfer and training initiatives developed in close collaboration with universities and technical centers. These exist in many countries; one example covered above comes from Chile’s copper sector. More generally, it appears that countries applying more inclusive approaches have been more successful than those following more protectionist methods focused on short-term goals and narrow targets. Policies that fail to consider long-term economic diversification objectives, focusing instead on narrow targets for the local sourcing of goods and services and the employment of the local workforce, are likely to generate businesses and skills that are dependent on the mining sector or even on a specific project (Weldegiorgis, Dietsche, and Franks, 2021; Lebdioui, 2019). Value added, created exclusively within the mining industry, perpetuates the vulnerability to commodity price fluctuations and macroeconomic shocks (Marcel et al., 2016; ACET, 2017). This Local content policies in the mining sector 71 is the case in particular of the downstream linkage policies implemented in Botswana. LCRs tend, incorrectly, to be considered fiscally neutral, not presenting any financial implications for the government. In fact, they can create fiscal shortcomings if they negatively affect the profitability of the mining sector, reducing the amount of taxes and royalties collected by the government. LCRs can be costly to implement due to the incentives granted to promote the use of local inputs and develop competitive local supplier businesses. Governments often face a trade-off between maximizing tax revenues and developing local content, but the extra cost of adopting LCRs can be justified only if it expands the tax base over the long term (Marcel et al., 2016). This will fail to materialize when LCRs reduce the industry’s competitiveness. Protectionist measures can lead to the prevalence of uncompetitive suppliers for long periods, raising production costs and having a detrimental effect on sectors using mining products as inputs (McCulloch et al., 2017). On the other hand, imported intermediate goods and services are an important channel to increase productivity and the adoption of new technologies and can play a significant role in integration in global value chains. By increasing the specialization in the production of specific inputs, they can generate economies of scale that maximize productivity and provide opportunities to move into higher value activities over time through upgrading (Korinek and Ramdoo, 2017). Finally, LCRs can cause disruption in global markets, in particular when they target minerals considered “critical” for certain key industrial sectors, inflicting supply risks, and affecting the sustainability of industries in resourcedependent countries. In a few instances in which LCRs are not WTO-compatible, this has led to costly disputes. For example, beginning in 2006, China imposed several restrictions on the export of rare earth metals and quotas on the export of unprocessed ores. This contributed to a steep increase in prices starting in 2010. In 2012, the United States initiated a dispute at the WTO against these restrictions, which the EU, Japan, and Canada joined as complainants. In 2014, the WTO Appellate Body decided in favor of the complainants, and China was required to remove its export restrictions. Policy implications: from local value added to integrating globalsupply chains Economy-wide impacts must be taken into account when considering implementing LCRs. If used, LCRs should be part of broader public policies, institutional arrangements, and partnerships that public authorities put in place to leverage the sector’s contribution to broader economic diversification (Dietsche, 2017). In particular, it should be kept in mind that supporting sectors that are dependent on mineral extraction may not increase economic diversification, may increase the impacts of resource price cycles, and may have substantial environmental consequences. 72 Jane Korinek and Paulo De Sa More generally, governments can adopt a broad set of horizontal measures to remove constraints to business development throughout the economy, with an aim to promote macroeconomic stability; provide regulatory clarity and stability; support SMEs; and improve infrastructure in areas like energy, transport, communications, information technology, and finance. A supportive environment can be created for the mining industry and its suppliers by removing overly burdensome regulatory requirements, improving business infrastructure, promoting skills development, strengthening institutional coordination and collaboration among local suppliers, and supporting increased access to finance, in particular for SMEs often ill-equipped to access global supply chains (AfDB and Bill and Melinda Gates Foundation, 2015). LCRs that are more likely to deliver on their expectations are (i) guided by a comprehensive understanding of firms’ procurement needs, strategies, and capabilities; (ii) based on a thorough understanding of the local capacities and bottlenecks to their enhancement; and (iii) cognizant of the factors that may impact policy effectiveness and the potential unintended consequences of such measures. If prescriptive LCRs are to be introduced, they should be based on forecasts of the mining industry’s future needs, spending projections for specific goods and services in existing and future mining projects, and assessments of the capacity of local firms to supply goods and services of sufficient quality at competitive prices. An evaluation of the skills and competencies required by the industry, as well as the timing and quantity of labor force requirements, should be used to inform potential policy reforms in education and training in order to strengthen and upgrade skills (IGF, 2018a, 2018c). LCRs must be aligned with what can be realistically implemented, given the capacity of local suppliers, to ensure the long-term competitiveness of the mining industry. Before the decision is made to introduce them, a detailed analysis of the procurement needs of mining firms should be undertaken, ideally in close conjunction with the firms, as was done by Canada and Australia. The analysis should assess the type of procurement opportunities that are available in major projects for each phase in their life cycle, as well as the capacity of domestic suppliers and the workforce to respond to the needs of these projects, leading to the identification of the gaps that need to be filled to enable them to take advantage of existing and potential opportunities. If demand-side policies are implemented, they tend to work better when pursued in conjunction with supply-side policies. For example, governments can complement regulation establishing minimum targets or quotas for local employees by implementing measures to promote training and skills development of the workforce. Although the form and content of training plans can be a voluntary component of a company’s human resources program, some countries have mandated training requirements or support for the development of training facilities (IGF, 2018e). Prescriptive LCRs entail long time frames between their announcement and effective implementation. As such, they are vulnerable to changes in market Local content policies in the mining sector 73 prices and in political commitment (ACET, 2017). If countries nonetheless want to implement them, they should be introduced gradually, in a phased approach, along with capacity-building efforts to allow domestic suppliers to adjust to them. Governments should periodically assess progress against objectives, adjusting them as local capacity increases and knowledge about future supply and demand improves. Any protection provided to local suppliers should be temporary and disbanded gradually. Sunset clauses, which prescribe that a regulation shall cease to have effect after a specified date, are good practice in local content policies (African Natural Resources Centre, 2021). Mining companies might be required or voluntarily take the initiative to design and implement SDPs. Larger companies are better placed to develop and implement such programs, but cooperation among firms, coordinated, for instance, by a Chamber of Mines, can facilitate the participation of smaller companies. Examples of successful supplier development programs include Anglo American’s Zimele Enterprise Program in South Africa and the joint Newmont–International Finance Corporation’s Ahafo Linkages program in Ghana. In some cases, mining companies support local suppliers by implementing measures other than LCRs, such as (i) breaking large contracts into smaller ones (unbundling) so that local firms may provide a smaller portion of the total contract tendered; (ii) posting all contracts, tender opportunities, and instructions for bidding processes on local supplier portals; (iii) sole-sourcing arrangements with local suppliers or firms from disadvantaged groups; (iv) stipulating requirements for outside suppliers to subcontract or enter into joint ventures with local firms; (v) assigning higher preference weightings to local businesses in competitive bidding processes or providing them with longer time frames for bidding; (vi) price-matching, allowing local suppliers to match the price of other suppliers; (vii) supporting local suppliers to obtain the certifications necessary to respond to tenders or compete for contracts; and (viii) using procurement methods in which bids are awarded to local suppliers when their price is within a certain percentage of the best offer – e.g., 2% in Ghana, 10% in Tanzania, or within 20% of the lowest foreign bid price in Kazakhstan (Esteves et al., 2014). Monitoring and reporting of policy outcomes, potentially with in-built sanctions for noncompliance, is key to increasing understanding about policy design and implementation. Governments and industry should publicly report on procurement processes, contracted suppliers, spending, and tax implications to improve oversight and accountability (Pitman and Toroskainen, 2020). Monitoring of LCRs has, in many cases, shown them to be ineffective. However, Australia and Ghana have implemented strong reporting mechanisms with sanctions if firms do not report on their local content objectives. Mining companies are highly dependent on contractors and suppliers as a source of technological innovation. Empirical evidence shows that local suppliers that thrive are generally incumbent firms whose experience in the market 80 Jane Korinek and Paulo De Sa p. 102312 www.sciencedirect.com/science/article/abs/pii/S0301420721003226? via%3Dihub White, S. 2017. Regulating for Local Content: Limitations of Legal and Regulatory Instruments in Promoting Small Scale Suppliers in Extractive Industries in Developing Economies. The Extractive Industries and Society. 4 (2). pp.260–266. https:// researchrepository.murdoch.edu.au/id/eprint/34791/ World Bank, and Kaiser Economic Development Partners. 2015. A Practical Guide to Increasing Mining Local Procurement in West Africa. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/21489 Yunis, J., and E. Aliakbari. 2020. Fraser Institute Annual Survey of Mining Companies 2020. Vancouver: Fraser Institute. www.fraserinstitute.org/sites/default/files/ annual-survey-of-mining-companies-2020.pdf Annex. Consistency with WTOrules Table 3.A1 Consistency of Local Content Policies with WTO Provisions Measures Relevant WTO Provisions Consistency with WTO Measures affecting sourcing of inputs Local procurement requirements Quota related to local sourcing A percentage of value added or specific volume of intermediate inputs to be purchased locally TRIMs illustrative list para. 1 (a) Quotas or specific percentages prohibited Trade balancing requirements Imports of one product linked to export performance of other products TRIMs illustrative list 1(b) for internal measures; 2 (a) for border measures Prohibited Manufacturing requirements Certain products are required to be manufactured locally TRIMs illustrative list Prohibited Limitations on imports Amount of goods and services that can be imported for the production process is limited GATT Art. III.5; GATT Art. XI.1; TRIMs illustrative list, para. 2(a) Prohibited Foreign exchange restrictions Restrict the inflow of foreign exchange attributable to an investor to constrain the amount of imported intermediate goods TRIMs illustrative list, para 2 (b)  Prohibited Exception for developing countries GATT Arts. XII and XVIII:B (Continued ) 82 Jane Korinek and Paulo De Sa Table 3.A1 (Continued) Measures Relevant WTO Provisions Consistency with WTO Preference for Investors to purchase local substitutes local substitutes for imports if “like product” is manufactured locally GATT Article III.4 (national treatment) if (i) imported products are accorded less favorable Prohibited treatment compared to local suppliers; (ii) imported goods and the domestic products are considered like products; and (iii) measures are inscribed in laws, regulations, and requirements. Ownership requirements Local equity Some proportion of participation equity must be held locally GATS Art. XVI for market access restrictions and Art. XVII for national treatment, in schedule of commitments Prohibited only if countries have taken commitments in their services schedules; otherwise not disciplined Employment requirements Local Specified employment employment targets must be met targets GATS Art. XVI for market access restrictions and Art. XVII for national treatment, provided in schedule of commitments Prohibited only if countries have taken commitments in their services schedules; otherwise not disciplined Local content policies in the mining sector 83 Measures Relevant WTO Provisions Consistency with WTO Quotas for A maximum number foreign of expatriate staff is employment specified National Certain staff must participation in be nationals or management a schedule for “indigenization” of management must be set Technology transfer requirements R&D Investors should requirements commit to invest in R&D locally Technology Specified foreign transfer technology be used locally Measures affecting production Minimum export Certain percentage of requirements production must be exported Trade balancing Imports must be a requirements certain proportion of locally produced exports, either in terms of volume or in terms of value Domestic sales Certain product may requirements not be exported Market reserve Some markets are policy reserved for local production GATS Art. IV; TRIPS Arts. 3, 7, and 8; SCM Agreement Arts. 2 and 8 GATT Art. III.5; GATT Art. XI.1; TRIMs illustrative list, para. 2(a) TRIMs illustrative list 1(b) for internal measures; 2 (a) for border measures GATT Art. III.5; GATT Art. XI: 1; TRIMs illustrative list 2(c) GATT Art. III.4 Prohibited only if countries have taken commitments in their services schedules; otherwise not disciplined Prohibited only if countries have taken commitments in their services schedules; otherwise not disciplined Prohibited Not disciplined Prohibited Prohibited Prohibited Prohibited (Continued ) 84 Jane Korinek and Paulo De Sa Measures Relevant WTO Provisions Consistency with WTO Product mandating requirements Some products to be exported by the hosting country only GATT Art. III.5; GATT Art. XI: 1; TRIMs illustrative list2(c) Prohibited Licensing requirements Investors to obtain license for production in the host country GATT Art. XI.1 Prohibited Technology transfers Investors are committed to a specified embodied technology TRIPS Arts. 3, 7, and 8; SCM Agreement Arts. 2 and 8 Disciplined Other measures relevant to LCPs State trading enterprises Foreign firms to enter into joint venture with SOEs Article XVII of GATT, applicable when SOEs enter commercial operations Provision does not regulate obligations of foreign firms to enter into joint venture with SOEs Subsidies to support local suppliers Governments give financial incentives to local suppliers to favor local products over imports SCM Art. 3.1(b) Actionable if specific; otherwise non-actionable Subsidies to R&D and innovation Government policies support R&D and innovation SCM Art. 8.2 Actionable if specific; otherwise non-actionable Art. = Article, GATS = General Agreement on Trade in Services, GATT = General Agreement on Tariffs and Trade, LCP = local content policy, R&D = research and development, SCM = Agreement on Subsidies and Countervailing Measures, SOE = state-owned enterprise, TRIMs = Trade-Related Investment Measures, TRIPS = Agreement on Trade-Related Aspects of Intellectual Property Rights, WTO = World Trade Organization. Note: Exceptions for developing countries – developing countries are permitted to retain TRIMs that constitute a violation of GATT Art. III or XI, provided the measures meet the conditions of GATT Art. XVIII, which allows specified derogation from the GATT provisions for the economic development needs of developing countries. Source: Korinek and Ramdoo (2017), adapted from Greenaway (1992); McCulloch et al. (2001). Table 3.A1 (Continued) Local content policies in the mining sector 85 Table 3.A2 Measures Not Prohibited by WTO Rules Measures Remarks Measures affecting imports Tariff measures WTO does not prohibit tariffs. Countries must bind their tariffs and can modify their tariff rates within the range if bound tariffs are different to applied tariffs. Nontariff measures (of a Generally prohibited (QRs, licensing, etc.) but quantitative nature) with the exception for imposition of import quotas for BOP purposes (Art. XVIII:B). This is temporary in nature. Measures to support enterprises Exchange rates No WTO agreement deals expressly with exchange rates, although GATT Art. XV concerns exchange arrangements. Government Permitted, except if a country is member of the procurement GPA. Export finance/ Allowed but may be considered an export subsidy insurance/guarantees if they are granted at premium rates insufficient to cover long-term operating costs and losses. Production subsidies Allowed if nonspecific* Trade finance Not prohibited Measures to promote technology Technology-related Not prohibited requirements for FDI (e.g., technological transfer) Support to R&D/ Unless specific, otherwise permitted innovation Human capital Not prohibited development Employment of local Not prohibited labor Regional assistance Not prohibited Investment incentives Export performance Not prohibited requirement as a condition for investment Equity requirement Not prohibited byFDI Measures subject to disciplines under specific circumstances Credit subsidies Not prohibited, provided they are not product or sector specific (Continued ) 86 Jane Korinek and Paulo De Sa Table 3.A1 (Continued) Measures Remarks Tax subsidies/holidays Not prohibited, provided they are not product or sector specific Clusters/EPZ/SEZ Not specially regulated by a particular WTO Agreement** but may be subject to disciplines when measures contravene other WTO disciplines (e.g., subsidies, etc.). Fiscal facilitation provided in SEZ is not prohibited. Contingency measures Safeguard measures These measures allow countries to apply import Anti-dumping measures restrictions in particular circumstances, provided Countervailing they can prove their economy/economic actors measures are affected by (i) a surge in imports (safeguard); (ii) a product that is being sold below its normal price on the domestic market by an exporting country (dumping); and (iii) a distorting effect of a subsidy by a foreign government. Art. = Article, BOP = balance of payments, FDI = foreign direct investment, GATT = General Agreement on Tariffs and Trade, GPA = Government Procurement Agreement, QR = quantitative restrictions, R&D = research and development, SCM = Agreement on Subsidies and Countervailing Measures, SEZ = Special Export Zone, WTO = World Trade Organization. * The WTO Agreement on Subsidies and Countervailing Measures disciplines the use of subsidies. The disciplines only apply to “specific subsidies,” that is, to subsidies available only to an enterprise, industry, group of enterprises, or group of industries in the country that gives the subsidies. They can refer to domestic or export subsidies. ** SEZ is mentioned in a footnote to GATT Art. XVI and in the SCM, excluding from the definition of a subsidy one of the fiscal facilitation measures provided to SEZs – an exemption from import duties and taxes on goods exported from SEZs. Source: Korinek and Ramdoo (2017). DOI: 10.4324/9781003415794-4 This chapter has been made available under a CC-BY-NC-ND 4.0 license. 1 Introduction When the Trump administration launched its revision of the treaty governing trade between the US and its neighbors, the US negotiators emphasized the need for stricter rules of origin. The US Trade Representative, Robert Lighthizer, reportedly asked his counterparts to raise the regional content requirement (RCR) to 85%, a large increase from the level set in 1993 (62.5%).1 Canada and Mexico balked at such a high rate, and the three parties finally settled on an increase to 75%, bolstered with additional binding requirements. The political appeal of stricter origin rules lies in the hope that they will increase domestic employment in the parts industry. Lighthizer (2020) acknowledged this intent, writing “The USMCA rebalances the NAFTA to promote increased production in the United States and North America.” From an economic standpoint it is hard to justify onerous restrictions on sourcing. If the goal is merely to limit imports of parts, then tariffs on parts would be a more efficient tool. While trade agreements that lack a common external tariff need some rule of origin to prevent back-door entry to the high-tariff market via the low-tariff country, this issue was not relevant in the USMCA negotiation for two reasons. First, because the actual differences in tariffs were small, so much smaller content restrictions would be sufficient to prevent this tariff-hopping.2 Second, it was the lower-tariff member, the US, that was asking for the stricter rules. Going back to the work of Grossman (1981), economists have investigated whether, even as protectionist devices, strict rules of origin could fail to achieve their goals. Grossman’s Proposition 3 (p. 591) states that small increases in local content requirements have ambiguous effects on industry value added, defined as the sum of value added in components and in final goods. Whereas the content protection policy causes an increase in the output of domestic components, it will normally result in a concomitant contraction of final good production. Which effect will dominate depends on how sensitive intermediate good production is to changes in its output price, and how sensitive final good production is to changes in the price of its intermediate input. 4 The unintended consequences of high regional content requirements Keith Head, Thierry Mayer, and Marc Melitz 88 Keith Head, Thierry Mayer, and Marc Melitz In this chapter, we extend the Grossman approach to take into account the very large number of diverse parts that go into modern manufactured goods such as automobiles. For each part, the firm decides whether to source it from inside the region (where there is a free trade agreement) or from outside countries. The core trade-off the firm faces is that within-region sourcing helps it comply with rules of origin (RoO), but necessitates forgoing opportunities to obtain cheaper parts elsewhere. In section 4, we give an overview of the theoretical model developed in Head et al. (2022) that analyzes these trade-offs. We show that RoOs generate competing incentives for part sourcing within a Regional Trade Area (RTA). Even though the rules are intended to relocate production of parts within the RTA, they can have the opposite effect when they are overly restrictive. This main result does not work via declines in final goods production, as in Grossman (1981). However, we also quantify the negative impact of higher costs induced by the RoOs for part production. This quantification exercise predicts how any given RoO would affect market share changes and the associated production and employment changes across all vehicle plans selling in the region. Drawing on the attractive aggregation properties of our model, we derive average price, market share, production, and part employment changes across groups of carlines – including the group of all carlines assembled within the region. This chapter is organized as follows. Section 2 provides an overview of recent changes in rules of origin that impacted the auto industry in North America and Europe. The following section presents empirical patterns of sourcing in North America that inform the model and the way we quantify it. Section 4 summarizes the key mechanisms of the model developed in Head etal. (2022). We then estimate the model to fit the pattern of sourcing observed at the level of individual car models prior to the 2020 changes in RCRs. Section 5 describes how we use that fitted model to evaluate the impact of counterfactual RoOs. Section 6 reports the effects of changing those rules for both NAFTA and the EU-UK trade agreement. 2 Changing rules of origin in North America and Europe Rules of origin in the auto industry were first introduced in the 1965 Auto Pact between Canada and the United States. To avoid non-US companies setting up sales enterprises in Canada to serve the US market, it was agreed that only cars with 50% content from the US and Canada would benefit from the new tariff-free regime.3 In the negotiation of the North American Free Trade Agreement in 1991, the American side sought a more restrictive rule. Irwin (2017) describes the initial negotiating positions and how they reached the peculiar regional content requirement of 62.5%: Rules of origin were particularly important in the case of automobiles. The US auto industry wanted high North American content rules to ensure that Mexico did not become an export platform for Japanese Unintended consequences of high regional content requirements 89 or other foreign producers who would simply send parts to Mexico for assembly and then ship the vehicles into the United States. .. . For NAFTA, the United Auto Workers pushed for an 80 percent rule, Ford and Chrysler 70 percent, and General Motors 60 percent. Mexico and Canada wanted to keep the 50 percent requirement in the US-Canada FTA, but reluctantly accepted 60 percent. US negotiators had promised auto producers a number higher than 60 percent to prevent their opposition. While they were able to persuade Mexico to go to 65 percent, Canada remained firm at 60 percent and so the negotiators split the difference and arrived at a 62.5 percent rule. Irwin goes on to describe how the US compromise led to an apoplectic call to the US trade negotiator from Ford’s CEO, who felt betrayed by the failure to obtain the promised 65%. The case points to the central importance assigned to rules of origin as well as the presumption that US producers would benefit from a stricter rule of origin than the one the US had settled on for NAFTA. When President Trump’s negotiators set out to replace NAFTA, one of their focal points was stricter rules of origin for the auto industry. Eventually, Canada, the US, and Mexico agreed in 2019 to replace the 1994–2020 NAFTA with a new agreement called the USMCA (in the United States). Lighthizer (2020) offered the following justification for stricter rules of origin: The USMCA rebalances the NAFTA to promote increased production in the United States and North America and to ensure that non-parties do not gain unwarranted benefits through the agreement. The USMCA features innovative rules of origin for automobiles and automobile parts that, once fully implemented, will create strong incentives to invest and manufacture in the United States and North America. The new agreement devoted 39 pages in an appendix to the new rules, so we cannot do full justice to their complexity here. The following were the main ways in which the requirements for qualifying for tariff-free treatment became more difficult for the auto sector: 1 The minimum North American regional content requirement (RCR) was increased to 75% (from 62.5%). 2 A new labor value content (LVC) rule requires that 40% to 45% of auto content be made by workers earning at least $16 per hour. 3 Seventy percent of both the steel and the aluminum going into each car must originate in North America. 4 Six “super-core” parts – including engines and transmissions – must themselves comply with the 75% RCR. The new requirements are clearly intended to discourage firms from sourcing parts from outside North America: if the vehicles currently 96 Keith Head, Thierry Mayer, and Marc Melitz 3.2 North American input cost shares (AALA data) Our source of data regarding variation in regional cost shares is based on annual reports mandated by the American Automobile Labeling Act (AALA) of 1992. The law requires that “A label with the US/Canada content percentage and related additional information must be displayed on these vehicles up to the time of first retail sale.” According to AALA, each new passenger motor vehicle must be labeled with the following information: 1 The percentage of US/Canadian equipment (parts) content 2 The name and percentage content for any countries other than the US and Canada that individually contributes 15% or more of the equipment content (with a maximum of two countries) 3 The countries of final assembly, engine manufacture, and transmission manufacture. The data are available in PDF form on the AALA website.8 Information on component suppliers other than the US and Canada begins in 2011. The cost share data is reported by AALA at the carline level, which usually corresponds to a brand-model assembled at a specific factory. AALA often provides more detail for carlines, with information such as engine size. We represent the model-level AALA data as a collection of cumulative densities in Figure 4.3. These are plotted with the original data pooled over the 2011–2020 period. We plot the CDFs separately for the cars that are the most 020406080 100 0.0 0.2 0.40.6 0. 81 .0 North American content share Fraction with < x J/K/G−made Canada RCR NAFTA RCR USMCA Figure 4.3 NAFTA Regional Cost Share by Location of Production (CDFs) Unintended consequences of high regional content requirements 97 potentially affected by the RoO, i.e., those produced in Canada, Mexico, and the US. We also present separate densities for the Japanese, Korean, and German brands that are produced in NAFTA (J/K/G make). Finally, we also plot a density for the models sold in the US but assembled in Japan, Korea, or Germany (J/K/G made). The AALA reports give estimates of the share of parts costs, not accounting for assembly costs. In order to compare those numbers to the RCR, we therefore need to add on the regional costs attributable to assembly. Figure 4.3 computes the overall regional cost share under the assumption that final assembly amounts to 15% of the total production cost of each regionally made car. Four main findings emerge. The majority of carlines in each NAFTA country have cost shares that indicate compliance with the 62.5% RCR prescribed by the original NAFTA. Second, compliance is highest in Canada, lowest in Mexico, and intermediate in the US. Car brands headquartered in the three major car-producing countries outside NAFTA have lower NAFTA input shares even when producing in NAFTA. Finally, North American cost shares for cars assembled outside North America tend to be very small. 4 A theoretical model of parts sourcing As we previously discussed, rules of origin (RoO) can generate competing incentives for the location of part production within a regional trade area (RTA). Those rules are intended to relocate the production of parts within the RTA; but when they are overly restrictive, the impact on regional sourcing will be reversed and part sourcing will be relocated outside the region. We now sketch a simple model based on our companion paper Head et al. (2022) that illustrates why RoOs will induce such a hump-shaped response for that regional part share. In order to focus on the sourcing decision for parts and the intuition for this hump-shaped response – which we call the Laffer curve for RoOs – we keep the location of assembly fixed. Our companion paper shows how RoOs will also impact that assembly location choice and how overly restrictive RoOs will not only lead to lower regional part sourcing but also induce final good producers to relocate assembly outside the region. 4.1 Model structure The potential for the downward-sloping segment of the RoO Laffer curve, where stricter RoOs lead to reductions in the regional part share, arises when final good firms (a carline producer in our data) make sourcing decisions for many parts. Although we would technically only need a minimum of two parts to highlight this effect, we develop a model with a continuum of parts due to its analytical tractability. And it also fits well with our empirical application in which car producers make sourcing decisions on a very large number of parts. 98 Keith Head, Thierry Mayer, and Marc Melitz Each car part can be sourced from either within the region at one cost or outside the region, denoted Foreign, at a different cost. Each part cost for regional and Foreign production is modeled as a stochastic draw from a Weibull distribution with parameter θ ≥ 1.9 We normalize the mean cost for regional production to 1. The mean cost of the Foreign-sourced parts is δ > 0. This parameter varies across firms. Firms with δ > 1 have a lower regional production cost for parts on average. As we mentioned earlier, we ignore the assembly location choice in order to focus on the part-sourcing decision (regional or Foreign); and we therefore do not model the associated assembly costs until the quantification in section 6. Free Trade (No Rules or Origin) When there are no RoOs, a firm δ decides whether to source each part from either within or outside the region based on whichever cost is lower. This is the firm’s unrestricted part-sourcing choice, which we denote with a subscript U. The resulting share of regionally sourced parts is given by the probability that the regional cost for a given part is lower than the Foreign cost. Given our distributional assumptions for the Weibull cost draws, that probability and resulting share is: dh hv U()sc dg -- 11 (1) Firms with higher δs have a comparative advantage in regional part production and hence source a higher share of their parts domestically. This sourcing decision then leads to a total parts cost (aggregating over both the regional and Foreign parts) of CU(δ) = χU(δ)1/θ. As we will see, these cost differences will be inconsequential for a firm’s response to a RoO, because that will only depend on how a RoO increases the firm’s cost above this benchmark CU(δ). Rules of Origin A RoO mandates that firms source a minimum fraction of their parts χR regionally, or else it will face a Most Favored Nation (MFN) tariff rate on the final good exported within the RTA. We model this additional cost as an average tariff τ > 1 incurred across all final good units produced. In the quantification in section 5, we will construct this average tariff rate based on the share of a carline’s within-RTA exports relative to all its other sales. If a firm chooses to comply with the RoO and avoid the tariff, it sources progressively more expensive parts regionally (relative to foreign-sourced) until the minimum threshold is met. In our companion paper, we show how the sourcing choices to comply with a RoO χR are equivalent to the ones the firm would make if a tariff were imposed on foreign parts (with the tariff revenue subsequently rebated back to the firm). We also describe the connections between a RoO specified as a regional part χR and a RoO specified as a regional cost share λR: a mandated minimum cost-share for regionally produced parts. Both types of RoOs have qualitatively identical effects on regional part sourcing Unintended consequences of high regional content requirements 99 because there is a monotonic relationship between χR and λR. This connection is important in the quantification because RoOs for cars in NAFTA and the EU-UK TCA are specified as cost shares. When a binding RoO χR > χU(δ) is mandated, the firm’s total part cost increases from CU(δ) to: CRRR dd d vv d dd d , vd cdf vd dd 11 1 (2) This represents an increase in the firm’s total part cost relative to its unrestricted (lower bound) cost CU(δ) given by the ratio  CCC RRU ff ff f ,,/() (0 = (0 >1 This cost ratio captures the compliance cost penalty associated with the RoO χR. It is represented in the top panel of Figure 4.4 as a function of the RoO χR for three different firms. Anticipating our empirical application, we use our fitted distribution for δs across NAFTA-assembled carlines. Firm 2 has δ2 = 0.12, which is the median δ (representing a 12% average cost advantage for NAFTA-produced parts).10 We then show two other firms (δ1 and δ3) that are, respectively, at the 5% and 95% percentile for that empirical distribution. For any given firm – a given δ – there is a range in which its unrestricted sourcing choice χU (δ) is above χR and therefore complies with the RoO. There is no cost associated with compliance, so  C (χR, δ) is at its lower-bound of 1. We denote this case compliant-unconstrained. As the RoO χR rises above χU (δ), compliance with the RoO entails a cost compliance penalty  C (χR, δ) > 1. As anticipated, this cost penalty then increases monotonically with the RoO χR: compliance becomes increasingly costly as the RoO becomes more restrictive. Looking across firms, we see that, as expected, the compliance cost with a given RoO χR is always higher for firms with lower δ whenever they are not unconstrained: those firms have a comparative advantage in Foreign-sourced parts, so complying with a given RoO is more expensive. 4.2 Compliance As we mentioned, a firm δ can choose not to satisfy the RoO χR and instead pay the average tariff τ. It will do so whenever the compliance cost is greater than the tariff penalty:  C (χR, δ) ≥ τ. In this case, we label the firm as non-compliant, and it then reverts to its unconstrained part sourcing with regional share χU (δ) and associated cost CU (δ) = χU (δ)1/θ. The horizontal line in the top panel of Figure 4.4 shows the example of a 6.2% tariff penalty. Continuing with our anticipated empirical application, this represents the non-compliance tariff that would be paid on average across all vehicles assembled in Mexico based on the empirical proportion of Mexican-assembled vehicles that are exported to 100 Keith Head, Thierry Mayer, and Marc Melitz Figure 4.4 Compliance Cost and Sourcing Decision for 3 Firms its NAFTA partners, the United States and Canada, and their associated MFN tariffs. The bottom panel of Figure 4.4 shows the regional part share chosen by the three firms, given their compliance decision. When the RoO χR is low enough, all three firms are compliant-unconstrained and choose their unrestricted part Unintended consequences of high regional content requirements 101 share χU (δ). This corresponds to the case of no compliance cost penalty,  C (χR, δ) = 1, in the top panel. As the RoO χR increases, firm 1, followed by firm 2 and then firm 3 become compliant-constrained: The compliance cost penalty  C (χR, δ) rises above 1 but remains below the tariff penalty τ. In this case, the firms choose the regional share χR to comply with the RoO. This is captured by the 45-degree increasing line in the bottom panel: a chosen regional share equal to the RoO. As the RoO χR further increases, firm 1 and then firm 2 choose non-compliance: the cost penalty is higher than the tariff penalty. In those cases, their chosen regional part-shares drop back to their initial unrestricted levels χU (δ). Note that firm 3 will never choose to be non-compliant: Complying with even the most restrictive RoO of 100% is still less costly than the tariff penalty. We label firms of this type as always-compliers. 4.3 Laffer curve for rules of origin Setting aside those firms that are always-compliers, we see in Figure 4.4 that increasing a RoO from 0% to 100% will initially induce firms to increase their regional part-share – when they are compliant-constrained – but will then induce those firms to sharply reduce their part-share once the RoO rises above a threshold where the firms choose non-compliance. In our companion paper, we show that this non-monotonic response, in this individual firm case an inverted-V, requires a firm-sourcing decision over multiple parts. When there is a single part, that non-monotonic sourcing response disappears: Increasing the RoO can never induce a firm to reduce its regional part-share. And we also show that as we smooth that inverted-V sourcing response at the firm level over a set of firms with heterogeneous δ, then the average regional sourcing share becomes a smooth inverted-U Laffer curve. So long as we exclude the always-compliers, then the average regional part-share returns to its initial (χR = 0) level as the RoO increases to its 100% upward bound. When we consider the full set of firms including always-compliers, then the average regional part-share remains above its initial level as the RoO increases to its upward bound.11 5 Simulating policy changes in the model The model delineated in the previous section provides key qualitative insights. Most importantly, it demonstrates the unintended consequences of an overly strict set of rules of origin. When the cost of compliance is higher than the penalty for non-compliance, firms will opt into non-compliance, cutting regional input use down to their unconstrained levels. The key unanswered questions are whether recent policy changes put North America into this range of counter-productive rules. Answering this question requires us to calibrate several different dimensions of heterogeneity. We do this by finding parameter values that induce the best fit between our simulated data and the observed data for the pre-USCMCA period, when the RCR was 62.5%. 102 Keith Head, Thierry Mayer, and Marc Melitz When taking the model to the data, we have to take a stand on the level at which the content decision is made. While the model refers to “firms,” the AALA reports show that different carlines owned by the same firm use very different shares of North American inputs. For example, the made-in-Mexico Ford Fiesta uses 80% North American parts, whereas the US-assembled Ford Mustang has 46% of its parts originating in North America. The Volkswagen Golf R, made in Germany, has only 1% of North American parts, but the Golf GTI assembled in Mexico has 42%. The US-assembled VW Passat has 61% for the version with a 2.0-liter engine (made in Mexico) and just 30% for the 3.6-liter version (engine imported from Germany).12 Thus, the data suggest that the content decision is taken in response to variation in relative costs (δ in the model) at the level of specific carlines. The actual decision-maker could be a plant manager or global headquarters. In the model, it does not matter whether the decision is centralized, because profit maximization implies that costs should be minimized for each carline. There is a single compliance decision for all the vehicles that come out of the same production line, regardless of their final destination. This assumption comes from observation in the IHS Markit data that it is extremely rare for the same carline to source a given engine or transmission from more than one country. Also, the AALA data provide single NAFTA shares for each carline. It is important to simulate the model at the carline level because the tariff penalty for non-compliance (τ in the model) varies greatly across carlines because of their different sales destinations. For example, the Ford Mustang has 2018 sales of 76,000 units in the US. These cars will not pay any tariff penalties for non-compliance with USMCA rules, nor will the roughly 12,000 units headed to Australia and China.13 Only the 7,600 Mustangs sold in Canada and the 1,900 sold in Mexico will face MFN tariffs as a penalty for noncompliance with the USMCA RoO. The situation of the Ford Fiesta made in Mexico is very different. The company sends the lion’s share of its total production (66,000 cars) to its USMCA partners: to the US (52,000 cars) and Canada (1,200 cars). Meanwhile, only 4,500 Fiestas stay in Mexico. The overwhelming dominance of export sales to NAFTA partners gives the Fiesta plant very strong compliance incentives, as compared to the Mustang. We capture this important source of heterogeneity by using the IHS Markit data to compute tariff penalties for every carline. The tariff penalty tends to be much lower than the MFN tariffs because large shares of output in the regional plants of a carmaker tend to stay within the country of production or go to markets outside the region (as in the Mustang example). Table 4.2 provides more granular information for the 20 largest tariff penalties. We use a simulation of our model to estimate the underlying heterogeneity parameters. The idea is that carlines receive their comparative advantage “draws” according to a particular “guess” for the mean and standard deviation of δ. At the same time, they draw a parameter determining the importance of assembly costs for that carline. Then the simulated carlines each decide Unintended consequences of high regional content requirements 103 Table 4.2 Top Tariff Penalties for USMCA Carlines in 2018 Brand Model Assembly country Tariff penalty sh. rest of RTA Chevrolet Silverado Mexico 1.23 0.96 Toyota Tacoma Mexico 1.22 0.97 Nissan NV200 Mexico 1.22 0.99 Ram 2500/3500 Mexico 1.21 0.94 Ram ProMaster Mexico 1.20 0.92 GMC Sierra Mexico 1.20 0.99 Ram 1500 Mexico 1.19 0.96 GMC Sierra Canada 1.18 0.80 Mercedes-Benz Sprinter United States 1.09 0.73 Chevrolet Silverado Canada 1.05 0.29 Volkswagen Golf SportWagen Mexico 1.03 0.96 Chevrolet Cruze Mexico 1.03 0.93 Nissan Note Mexico 1.03 0.86 Volkswagen Golf Mexico 1.03 0.81 GMC Terrain Mexico 1.03 0.98 Toyota Corolla Canada 1.03 0.85 Infiniti QX50 Mexico 1.03 0.93 Buick Regal Canada 1.03 0.97 Dodge Journey Mexico 1.03 0.94 Dodge Charger Canada 1.03 0.89 Note: Head et al. (2022) provides the formula used to compute the carline-level tariff penalty in a way that takes into account market share changes in response to tariff changes. whether to comply with a content requirement of 62.5%. Depending on the assembly cost share, this RCR converts to a particular parts costs share (λR in the model), which in turn converts to an implied share of regional parts (χR in the model). If compliance is too costly relative to the tariff penalty, then the carline selects its unconstrained cost, minimizing North American parts share. The result is a vector of parts costs shares emerging from the simulated model. Recognizing that the model is an approximation, and the data reporting in AALA is far from perfect, the simulation builds in random measurement error.14 The result is a simulation-based distribution of North American parts shares, which we compare to the actual distribution from the AALA reports. We quantify the discrepancy in terms of the sum of squared deviations between model and data. The algorithm then repeats the procedure for a large grid of different guesses for the parameters, selecting the ones that achieve the best fit between simulation and observation. Head et al. (2022) provides a more formal description of this procedure for estimating the model parameters. The estimated parameters allow the distribution of the simulated carlines to tightly fit the distribution of North American content reported by AALA. To provide external validation for the quantified version of the model, we follow the common practice of considering a feature in the data that was not part of the original moment-matching exercise. For this purpose, we compare the implied RoO compliance rates (also referred to as preference utilization 104 Keith Head, Thierry Mayer, and Marc Melitz rates) for auto trade (HS 8703) to those that emerge from the simulation based on the calibration described earlier. As shown in Table 4.1, the true rate of preference utilization for US-made cars entering Canada was 97% in 2019 (before the change in the regional content requirement in 2020). The calibrated model obtains a rate of 92%. Thus, our model is able to closely mirror the distribution of North American content rates at the carline level and also match reasonably well the RoO satisfaction rates observed for aggregate trade flows within North America. After obtaining the best-fit values, we can solve the model for any potential RCR. This requires computing how each individual carline will respond to a stricter RCR. Depending on their parameter draws, they might increase regional parts shares just enough to match the new requirement, or they might opt into non-compliance. Based on this decision, the change in costs (from increasing regional content in response to a stricter rule) or the tariff penalties (from opting not to comply with a stricter rule) will reallocate market share towards foreign carlines, as well as those domestic carlines that were not complying before the stricter rule. In computing the changes in this step, we take advantage of the aggregation properties of the constant elasticity of substitution (CES) demand system. This provides an exact aggregation for the resulting changes in the price index and employment in the next section. 6 Quantification of the impact of RoO changes In this section, we use our model, with parameters chosen to fit the distribution of regional content by North American carlines, to quantify the effects of two recent changes in RoOs. The first is the tightening of RoOs for North American vehicle trade, which was one of the most salient features of the USMCA. The second is the application of rules of origin to UK–EU trade, required by Britain’s exit from the customs union in the final Brexit deal. We evaluate changes in the strictness of the RoO, as measured by changes in the RCR for the enacted policies. We also consider alternative RCR levels that might have been chosen. For each policy change, we report outcomes for groups of carlines based on their compliance decisions before and after the RoO changes. For example, the first group in each table is the one for carlines that comply exactly with the old RoO but then decline to comply with the new RoO. The first numerical column shows the share of carlines in each group (in percent). The last four columns report the simulated changes induced by the change in the RCR. These outcome variables comprise the percentage changes in the price index, the group’s market share, the weighted average regional parts share, and employment. 6.1 USMCA Table 4.3 describes the simulated outcomes for the USMCA increase in the RCR from 62.5% to 75%. According to the calibrated model, just over a third of carlines switch from complying unconstrained to complying at the Unintended consequences of high regional content requirements 105 Table 4.3 Increase in RCR from NAFTA (62.5%) to USMCA (75%) Compliance status under: Percent changes in NAFTA USMCA Share of Price Mkt Parts Parts (RCR = 62.5%) (RCR = 75%) carlines share share Emp. ComplyNon-compliant 16.90 0.57 −1.05 −10.40 −11.85 constrained ComplyNon-compliant 7.10 0.27 −0.16 0.02 −0.40 unconstrained ComplyComply7.30 1.32 −3.23 20.97 15.53 constrained constrained ComplyComply34.50 0.21 0.00 8.26 8.03 unconstrained constrained Non-compliant Non-compliant 8.30 0.00 0.65 0.00 0.65 ComplyComply25.80 0.00 0.65 0.00 0.65 unconstrained unconstrained All All 100.00 0.28 −0.20 2.80 2.30 Notes: “Share of carlines” refers to the percentage of all domestic carlines in the corresponding status tuple. “Parts share” is a quantity-weighted average of the shares of parts from NAFTA origins across regionally assembled carlines. “Parts Emp.” is employment in parts manufacture for domestically assembled vehicles. minimum required level of 75%. These carlines will increase their regional parts shares by about 8%. The increase in average costs for the group is just one fifth of a percent. There is no discernible reduction in market share for this group, and its employment rises by almost the same amount as its average parts shares. Greater employment gains are recorded by the 7.3% of carlines that were just complying at 62.5% and raise their regional content up to 75%. These carlines increase their parts shares (X) by 21%, slightly more than the overall cost change of 0.75/0.625 − 1 = 20%. The implied rise in employment is just under 16%. The dampening comes from the 3% market share reduction for this group, which itself follows from their 1.32% rise in their average price. The increase in employment for the constrained compliers is mostly offset by a reduction in employment by carlines that stop complying once faced with the 75% RCR. The overall employment gain is just 2.3%, much lower than the naive expectation of 20% (0.75/0.625 = 1.2) that would follow from assuming that all carlines mechanically comply with the RoO. While the employment gains are modest, so are the price increases faced by consumers: the price index for regionally assembled cars rises by just 0.28%. As predicted by the convex cost curves shown in Figure 4.4, there will be a higher cost of further rises in the RCR. Table 4.4 reports the results of a counterfactual rise in the RCR from 75% to 85% (the original US ask during the USMCA negotiations). The last row of the rightmost column gives an interesting message for policy. It shows that had the US succeeded in negotiating an 85% RCR, this would have reduced 112 Keith Head, Thierry Mayer, and Marc Melitz 7 Policy implications and discussion The USMCA was welcomed by the chief lobbyist for Canadian auto parts manufacturers, Flavio Volpe. In an interview, he contended, “That deal [USMCA] ... is the best single positive hit for supplier business across North America in the history of the auto business. We think there’s going to be 25% more in absolute volume bought from local suppliers.” In addition, the head of the Mexican auto parts industry association predicted a ten percent increase in production in Mexico’s part sector.19 In contrast, the calibrated version of our model implies a much smaller effect of 2.3% (Table 4.3, bottom row). What is it about our model that implies much lower employment gains from RoO increases than naive calculations? The key point is that complying with a strict rule of origin is a choice. The benefit is preferential tariff access to the other North American markets. However, so long as the US maintains its 2.5% MFN tariff on finished cars, this is not a huge penalty. Moreover, some German factories in the US may care far more about their sales in other markets – such as China, for example – than they do about losing sales in Mexico or Canada. If bringing transmission sourcing to North America will add to the costs and make the vehicle non-competitive in China, the firm might prefer not to comply on sales to Mexico or Canada and then source engines from Europe as well if the only reason it had only sourced locally was to comply with the old NAFTA rules. The results from our quantification suggest that the old NAFTA rule and the current TCA rule are both under the parts employment-maximizing levels. However, the original Trump administration demand of 85% would have been counter-productive even from a purely protectionist standpoint. Our results also suggest the 100% content requirements for batteries for EVs are likely to lower employment while significantly raising the costs of EV adoption. Notes 1 Husisian et al. (2018) note the 85% proposal in their overview of the USMCA. 2 Felbermayr et al. (2019) present evidence that this argument applies to most rules of origin. 3 Anastakis (2005) provides a book-length treatment of this pioneering regional agreement. 4 “Inside the Brexit deal: the agreement and the aftermath” George Parker, Peter Foster, Sam Fleming and Jim Brunsden, Financial Times January 21, 2021. 5 “What’s driving the EU on rules of origin?” Jim Brunsden, Financial Times October 29, 2020. 6 By contrast in the main manufacturing countries outside North America – Japan, Korea, and Germany – the USA-USA pairing is used for just 1% of cars. 7 Figure 4.2 applies the great circle formula to calculate the distance between engine (or transmission) factories and the final vehicle assembly factory. Since engines and transmissions are too heavy and bulky for air shipment, road, rail, or sea distances would be more accurate. Past work finds high correlations between great circle and actual road distances within countries. For intercontinental trade, air routes diverge in a more severe way from sea routes. Thus, we should expect that any Unintended consequences of high regional content requirements 113 measurement error is larger for long distances, but we see relatively little trade at distances over 2,000 km. 8 www.nhtsa.gov/part-583-american-automobile-labeling-act-reports 9 The parameter θ governs the variance of the cost draws. As θ increases, the variance decreases. In the limit, as θ goes to infinity, the variance goes to zero, and there is no variation in the cost draws around their mean. 10 We also set θ = 4. 11 Hypothetically, if the distribution of δs is such that it is dominated by alwayscompliers, then it is possible for the average regional part-share to monotonically increase with the RoO. However, we show that this is not the case for NAFTA. 12 All these percentages are cost shares from the 2019 AALA report. 13 The tariffs China imposes on US exports do not depend on their North American content. 14 Among the sources of error are the AALA exemption for reporting Mexico content if it is below 15%. Additional measurement error comes from rounding, which the law permits to the nearest 5%. We also intend for the error to capture deviations from the continuum assumption in the model. Since many parts have non-negligible cost shares, a firm that intends to “just comply” will in fact be observed to over-comply depending on the share of the last part. 15 The Verge, August 8, 2022 16 Business Insider August 10, 2022 17 Renault, Peugeot, Seat, and Skoda are examples of popular brands in Europe that are not offered in the US. 18 There are some complexities in the UK-EU TCA as regards electric vehicles. 19 Reuters, October 1, 2018. References Anastakis, D. (2005). Auto Pact: Creating a Borderless North American Auto Industry 1960–1971. Toronto: University of Toronto Press. Felbermayr, G., F. Teti, and E. Yalcin (2019). Rules of origin and the profitability of trade deflection. Journal of International Economics 121, 1032–1048. Grossman, G. M. (1981). The theory of domestic content protection and content preference. The Quarterly Journal of Economics 96(4), 583–603. Head, K., T. Mayer, and M. Melitz (2022). The Laffer curve for rules of origin. Manuscript. Husisian, G., A. Gomez-Strozzi, and A. Alvarez (2018). International trade: A new dawn for North American trade. Technical report, Foley LLP. Irwin, D. A. (2017). Clashing Over Commerce. Chicago: University of Chicago Press. Lighthizer, R. E. (2020). 2020 Trade Policy Agenda and 2019 Annual Report. Washington, DC: Office of the US Trade Representative. DOI: 10.4324/9781003415794-5 This chapter has been made available under a CC-BY-NC-ND 4.0 license. 1 Introduction Due to the rapid development of global value chains (GVCs) over the last three decades, the “Made in” label typically applied to manufactured goods, attributing them to a specific economy, has become an archaic symbol, as most manufactured products are now “Made in the World” (they are produced at stages in several countries, with value added at each stage). The rise of GVCs has significantly changed the nature and structure of international trade and investment and brought considerable benefits to China and is the major driver behind China’s rapid industrialization. However, the growth of GVCs has slowed since 2012, after a quick recovery following the Global Financial Crisis (World Trade Organization, 2019). Bakas (2019) points to this decline as suggesting that the world has entered a period of “slobalization.” The GVC participation rate in China has plateaued since 2007 and was below the world average in 2019 (Asian Development Bank, 2021). The trade war between the US and China in 2018 further worsened China’s international environment. The US government first attempted to reduce US imports from China through higher tariffs, then proceeded to impose strict export controls to cut off key high-tech components supply to Chinese hightech firms such as the ban of semiconductor sales to Huawei. For some hawkish members of the US Congress, undoing 40 years of ever-closer economic relations with China and rolling back US reliance on Chinese factories was always one of their political objectives. An “Economic Prosperity Network” of likeminded countries, a concept initially proposed by the Trump administration and inherited and strengthened by the Biden administration, aims to convince Western firms to extricate themselves from China and instead partner with firms headquartered within member countries of the network based on the common goal of reducing economic dependence on Beijing (“friend shoring”). Economic nationalism and various protective measures are on the rise. China has made strategic moves to prepare itself for this less favorable international economic environment. Beijing has announced a dual-circulation economic strategy that emphasizes domestic consumption as the major vehicle for economic development. Since Beijing launched a campaign to develop 5 LCR Policies in China and their impacts on domestic value added in exports Kun Cai and Zhi Wang LCR policies in China 115 more advanced technologies at home and rely less on the United States and other Western suppliers in 2012, it has been pursuing its own form of “made in China” for more than a decade. Achieving technological independence from the West, especially the United States, has been a stated goal of the Chinese government and reaffirmed by the current leader.1 Beijing is pursuing two key objectives: (1) eliminating its dependence on foreign countries for critical technologies and products and (2) encouraging domestic indigenous firms to bolster their own capacity for innovation in order to become leaders in advanced technologies. To achieve such key objectives, various government agencies at different levels in China proposed and implemented a series of industrial policies, including some implicit local content requirements (LCRs), to encourage domestic production and innovation. This chapter reviews these policies and measures the changing trend of domestic content in China’s exports from 2007 to 2017 based on detailed trade statistics from the China Custom administration and the most recent national input-output tables (IOTs) published by National Bureau of Statistics (NBS). We also seek to assess the implications of various implicit LCR measures proposed in China on domestic content in Chinese exports. The chapter is organized as follows. Section 2 reviews the major industrial policies and implicit LCR measures in recent years proposed by Chinese central and local governments and by major manufacturing industries. Following Koopman et al. (2012), Section 3 outlines the conceptual framework for estimating domestic value added (DVA) in a country’s exports when processing exports are important. We extend their methodology to decompose production activities into pure domestic, traditional trade and GVC activities at the country/sector level based on national IOTs. Section 4 presents the major empirical results and uses them to evaluate the impact of China’s LCR measures on domestic content in Chinese exports. We find no empirical evidence that those policy measures implemented by the Chinese government in recent years have played any significant role in promoting domestic content in its exports, at least at the aggregate level. Section 5 concludes. 2 Recent local content requirement policy development inChina2 A local content requirement is a measure that supports the use of local inputs in the production of goods or services as a precondition for gaining market access or obtaining financial incentives. Countries that have taken these measures hope to compel foreign companies to source from local firms to promote the development of their own industries. This support of local inputs incentivizes firms to select their suppliers based on their nationality rather than quality and cost. The scope of LCR measures is not clearly defined. Hestermeyer and Nielsen (2014) classified LCR policies into three categories: licensing, government procurement and financial incentives. Hufbauer et al. (2013) believe LCRs can take many forms, including price preferences awarded to domestic firms that bid on government procurement contracts, mandatory minimum 116 Kun Cai and Zhi Wang percentages required for domestic goods and services used in production, import licensing procedures designed to discourage foreign suppliers and discretionary guidelines that both encourage domestic firms and discourage foreign firms. They identified 117 LCR measures implemented across the world since the 2008 financial crisis and pointed out the distinctive characteristics that LCRs have compared with other trade policies. Following their work, Stone et al. (2015) provide a quantitative analysis of the localization barriers to trade. They group various LCR measures in two dimensions, which are the targeted market and identified benefits. The targeted markets include inputs, ownership, labor, government procurement and data. The benefits include market access, price preference, tax policies, government funds and domestic branding schemes. The OECD had deemed recent “Made in XX” or “Buy XX” programs initiated by some countries as localization barriers to trade.3 The LCR measures applied by China before its WTO accession were explicit. For instance, preferential tariff and tax incentives were provided based on the percentage of local inputs, and foreign enterprises were forced to follow the mandatory technology transfer requirements. After 2001, explicit LCR percentages for goods or services were gradually lifted. However, implicit localization trade barriers ingrained in the implementation of industrial policies emerged. These implicit LCRs aim to promote the innovative capacity of China and to cultivate indigenous domestic companies. On the surface, these policies treat producers equally regardless of nationality, while in practice, foreign producers may be encouraged to conduct localization strategies voluntarily, or it may be the case that only indigenous firms truly benefit from these preferential policies. Due to their opacity and covertness, these LCRs can be difficult to identify. The heterogeneity of localization policies is prominent across sectors. A sector-by-sector approach is taken to present the localization policies in China, which take the forms of market access, subsidies, licensing and government procurement. We focus on the automobile, integrated circuits (IC), telecommunications, pharmaceutical and medical equipment industries because they are of vital importance to China’s industrial system, and for that in Made in China 2025, specific targets of the localization rate or market share are set forth for many critical materials, products and processing equipment in these industries. By looking at the LCR policies in these sectors, we hope to shed some light on the roles they played in production activities and their effect on the changing trend of domestic contents in Chinese exports. 2.1 Auto industry The auto industry is one of the pillar industries in the Chinese economy. The value added of auto vehicle and auto parts production accounts for approximately 2% of China’s GDP.4 In 2021, both the domestic sales and production of cars in China reached 26 million, which means that approximately 32.5% of total world auto production is conducted in China and 31.8% of global automobile sales are conducted in China, making China the largest car manufacturing and consumption country. LCR policies in China 117 Auto manufacturing in China relies heavily on locally made components. Figure 5.1 presents the value of the imported auto parts per vehicle and - 5,000 10,000 15,000 20,000 25,000 30,000 - 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 Numbers of cars produced (right axis)Imported auto parts per car (left axis) thousands of cars Imported auto parts per car produced (dollars) Figure 5.1 Impor ted Automobile Parts and Automobile Production of Major Producing Countries in 2020 Notes: Auto production includes both commercial and passenger vehicles. Imported auto parts per car are calculated as total imported auto parts divided by the number of cars produced in a country. Sources: Authors’ calculation based on production statistics from the International Organization of Motor Vehicle Manufacturers (OICA). https://www.oica.net/production-statistics/. UN Comtrade Database. https://comtrade.un.org/data/. Table 5.1 MNEs and Their Joint Ventures in China MNEs Country Joint ventures Year of Shareholdings in China establishment Volkswagen Germany SAIC 1984 50% by Shanghai Auto Volkswagen Industry Co. (SAIC), 40% by Volkswagen, 10% by Volkswagen China FAW 1991 60% by China FAW Group Volkswagen Co., (FAW), 20% by Volkswagen, 10% by Audi AG, 10% by VW China JAC 2017 50% by Volkswagen, 50% Volkswagen by JAC Group (JAC) In 2020, VW brought in 50% stakes of JAC Group and increased the share in JAC VW to75%. Daimler Beijing 1983 51% by Beijing Automobile Benz Co., LTD (BAIC), 49% by Daimler Foton 2011 50% by Beijing Foton, 50% Daimler by Daimler (Continued) 118 Kun Cai and Zhi Wang Table 5.1 (Continued) MNEs Country Joint ventures Year of Shareholdings in China establishment Fujian 2007 35% by Beijing Auto, 15% Daimler by Fujian Auto, 50% by Daimler BMW BMW 2003 50% by BMW, 50% by Brilliance Auto. In 2022, BMW increased stakes to 75%. Hyundai Korea Beijing 2002 50% by Beijing Auto, 50% Hyundai by Hyundai Toyota Japan FAW 2003 35% by FAW, 50% by Toyota Toyota, 15% by Tianjin FAW Xiali Co. LTD The current ratio is 38% by FAW, 32% by Toyota, 30% by Tianjin FAW Toyota (TFTM). GAC 2004 50% by Guangzhou Toyota Automobile Group Co. LTD (GAC), 30.5% by Toyota, 19.5% by Toyota China Honda Guangqi 1998 50% by GAC, 40% by Honda Honda, 10% by Honda Technical Research Industry (China) Investment Dongfeng 2003 50% by Dongfeng Motor Honda Group Co. LTD (DFG), 40% by Honda, 10% by Honda China Nissan Dongfeng 2003 50% by DFG, 50% by Nissan Group of China GM US SAIC GM 1997 50% by SAIC, 50% by GM SAIC-GM2002 50.1% by SAIC, 44% by Wuling GM, 5.9% by Guangxi Auto (formerly called Wuling) SGM 2004 25% by SAIC, 25% by GM Norsom China, 50% by Shanghai GM Ford Changan 2001 50% by Changan, 35% by Ford Ford Asia Pacific Motor Holdings LTD, 15% by Ford China JMC Ford 1997 41% by Nanchang Jiangling Investment Co., LTD, 32% by Ford, others by public shareholders Tesla – 2018 100% by Tesla Source: Collected from official websites of the companies in the table and the enterprises big data platform operated by Baidu, https://aiqicha.baidu.com/. LCR policies in China 119 the number of cars produced in major car-manufacturing countries. Each car produced in China used approximately $1,021 of imported components, far less than Canada ($14,564), France ($9,792), Germany ($9,200) and the US ($6,806). Other emerging economies, such as Mexico ($6,901), Russia ($5,330) and Brazil ($2,615), also use a higher value of imported parts than China. Among the top 15 auto-manufacturing countries, Japan has the lowest value sourced from overseas, approximately $778. The imported value per car in Korea ($1,154) and India ($961) is roughly the same as that in China. One reason that auto components and parts remain largely locally sourced is related to the auto market access policy in China. In 1994, the Automotive Industrial Policy was published, which requires foreign carmakers to form a joint venture (JV) to gain access to the Chinese market. An equity cap of 50% is also set for foreign shareholders. Table 5.1 lists the joint ventures and the shareholding of multinational enterprises (MNEs) in China. The requirements of JVs and the equity caps have gradually relaxed in recent years. BMW raised its equity share to 75%, becoming the first foreign company to have a majority share in auto JVs in China. In 2019, Tesla became the first foreign automobile company with a wholly owned Gigafactory in China, as shown in Table 5.1. In recent years, fiscal subsidies have been provided to new-energy vehicle (NEV) producers, which is another one of China’s efforts to promote domestic auto production. The subsidies are not exclusively provided to indigenous manufacturers, but domestic producers seem to have benefitted the most. Table 5.2 shows the top 20 companies that received the most fiscal subsidies from 2017 to 2020. Each year, they collectively account for approximately 90% of total subsidies to NEVs. In 2021, China published the Provisions on the Security Regulation of Automobile Data, laying down the rules that all important data must be stored within the geographic boundaries of China. The so-called important data include data of pedestrian flow, traffic stream, data of electric automobile charging networks, videos and images of car plates and drivers’ faces. Any cross-border transfer of such data must be examined by local authorities. Complying with the regulations on auto data, many MNEs, such as Tesla, have set up data centers5 in China to facilitate the localization of data storage and processing. 2.2 Integrated circuit (IC) industry The goal of IC-promoting schemes in China is to cultivate comprehensive and mature domestic IC supply chains, covering all production stages and achieving chip self-sufficiency. Statistics show that the global production share of China’s IC industry was only approximately 5% in 2020, far behind that of the US (47%), Korea (20%), Japan (10%) and Europe (10%). Additionally, China is more concentrated on relatively capital-intensive activities, including materials, wafer fabrication, assembly, packing and testing. The value added captured by China accounts for approximately 9% of the total semiconductor value chains.6 120 Kun Cai and Zhi Wang Subsidies , million) 1448.91 776.33 549.93 496.38 275.74 232.18 190.75 187.98 170.9 136.93 135.53 122.72 118.87 105.32 100.47 82.22 79.06 75.06 68.46 63.56 (CNY 5890.34 e t_99ae13aa81d04a358 2017 Company BAIC Motor DFAC Geely Auto y MotorCher Zhongtong Bus Dayun Auto Guangtong Manufactur Changan Auto BYD Auto Sunlong Bus Golden Dragon Bus V SAIC Motor Jiangnan Auto CRRC Times E Shuchi Bus Foton Auto otal Yaxing Bue Joylong Auto y AutoictorV Xinchufeng Auto t/2021/ar T Subsidies , million) 696.06 636.35 468.71 369.93 341.56 340.52 253.3 (CNY 465.22 245.46 227.23 220.93 215.56 187.71 152.81 112.98 109.16 92.38 91.32 83.63 63.57 6075.8 .cn/zwgk/wjgs/ar eManufactur V miit.gov way Motors all Motor ://wap. 2018 Company DFAC BAIC Motor Geely Auto BYD Auto JMC Motor y MotorCher SAIC Motor Changan Auto Jiangnan Auto Guangtong JAC Motor Golden Dragon Bus inner Sunlong Bus BMW CRRC Times E otal W Yutong Bus eat WGr Zhongtong Bus Foton Auto . https T echnology Subsidies , (CNY million) 1134.74 966.35 890.39 801.73 783.86 741.16 599.59 343.26 315.32 275.76 267.54 266.12 175.32 166.04 161.01 159.81 142.51 128.95 121.79 91.06 9578.99 mation T , 2017–2020 BAIC Motor Golden Dragon Bus V Geely Auto SAIC Motor Gr CRRC Times E y Motor BYD Auto Sunlong Bus -VW 2019 Company utong BusY DFAC Zhongtong Bus Changan Auto all Motoreat W Cher SAIC Foton Auto BMW Anhui Ankai Auto way Motors y and Infor V Subsidies Beneficiaries inner axing Bus otal W GAC Motor Y T y of Industr Subsidies , million) 2123.37 2046.87 951.32 830.56 550.41 539.57 410.86 379.02 338.97 303.97 300.87 248.74 238.44 195.62 166.74 150.89 110.42 105.16 80.59 66.6 (CNY 10537.17 om the Ministr op 20 NE .html T all Motor oyota 5.2 esla Shanghai GAC Motor y Motor Based on data fr 2020 Company BYD Auto eat W T Gr Cher JAC Motor SAIC Motor Yutong Bus Geely Auto BAIC Motor SGMW Auto Xiaopeng WM Motor Haima Motor T able - T GAC Golden Dragon Bus Hozonauto Zhongtong Bus DFAC Lixiang Auto : otal ce T Sour e0443700fe100ce LCR policies in China 121 Table 5.3 Invested Firms and the Shareholdings of National IC Fund Phase I (accessed on April 24, 2022) Segmented production Firms Shareholdings of ICF Subscribed capital (CNY million) Wafer fabrication Semiconductor Manufacturing North China (Beijing) Corporation (SMNC) 32.00 10291.20* Semiconductor Manufacturing South China Corporation (SMSC) 14.56 6341.55* Huahong Wuxi 29.00 3698.40* Yangtze Memory 24.09 13558.42 Shanghai Huali Integrated Circuit Corporation 39.19 11600 Ningbo Semiconductor International Corporation (NSI) 13.55 60 Beijing Yan Dong Microelectronic Technology Co., Ltd. (YDME) 18.84 113 Silex MiroSystems 30.00 600 Compound semiconductor Sanan Optoelectronics 6.47 - Beijing Century Golden Light Semiconductor Co., Ltd. (CENGOL) 9.61 29.57 Packing and testing JCET 13.31 - Tongfu Microelectronics Co., Ltd. (TFME) 15.13 - JCET Shaoxing 26.00 1300 Specialized equipment and key components NAURA Technology Group Co., Ltd. (NAURA) 7.48 - Hangzhou Changchuan Tech (CCTECH) 6.76 - Shanghai Precision Measurement Semiconductor Technology, Inc. (PMISH) 7.30 100 RSIC scientific instrument (Shanghai) Co., Ltd. 8.78 37.58 Sky Technology Development Co., Ltd. Chinese Academy of Sciences (SKY) 19.73 - National Silicon Industry Group (NSIG) 20.84 - Materials Jiangsu Xinhua Semiconductor Material Technology (Xinhua Semiconductor) 23.56 306.29 BDStar Navigation 8.57 - Chip design Unisoc (Shanghai) Technologies (Unisoc) 13.96 705.88 Shanghai AisinoChip Electronics Technology Co., Ltd. (AisinoChip) 21.06 13.5 Empyrean Technology 11.1 48.19 Telink Microelectromics Shanghai (Telink) 11.94 21.49 Note: * indicates the original figures are in USD and the numbers are converted using the foreign exchange rate of 6.70 RMB/USD. — indicates data not available. Source: Based on the financial statement of public companies; the enterprise big data platform operated by Baidu, https://aiqicha.baidu.com/. 128 Kun Cai and Zhi Wang activities of traditional trade. The last part is value added embodied in exports of intermediate goods and services, which will be used in production activities in other countries and involves cross-country production sharing, so it can be defined as GVC activities. It includes both the simple GVC (DVA crosses borders only once) and complex GVC (DVA crosses borders multiple times) activities defined in Wang et al. (2017). The DVA decomposition results from Equation (5) will be the same based on a national IO table as using a global ICIO table if there is no need to separate the part of DVA that is first exported but eventually returns home. It is also straightforward and computes domestic contents in a country’s gross exports as the sum of the second and the third part of Equation (5). It is important to note that this method is limited because it relies on national IOTs. Although it can decompose production into pure domestic, traditional trade and GVC activities and compute DVA in exports, including direct and indirect value-added exports via upstream or downstream sectors, it cannot estimate DVA first exported but returned home via imports, indirect exports of DVA via third countries, distinguish simple and complex GVC activities, decompose bilateral trade flows, measure double counting due to intermediate goods trade and trace foreign value added and/or vertical specialization by country sources. These quantitative measures must be obtained from ICIO tables. 3.2 Domestic content in exports – when processing exports12 are important The production of processing exports often has a very different intensity in the use of imported inputs than that in domestic final sales and normal exports. To reflect such heterogeneity, one needs to keep track separately of the IO coefficients of the processing exports and those of domestic final sales and normal exports. The extended IO table with a separate account for processing exports is represented by Table 5.7. In such an extended IO table, domestic production has been separated into two parts: (1) production for domestic final demand and normal exports represented by superscript D and (2) production of processing exports represented by superscript P. Mathematically, the model can be specified as Equations (6) and (7), IA A I XE E YE E DD DP P P DP P dd f g g g j l | | d f g g g j l | |=d f g g g j l | | 0 (6) AXEAEY M MD PMPP M - 90 === (7) where A DD = Z DD ( XE P −− ) 1 and A DP = Z DP (EP)− 1 denote the input coefficient matrix for the production of domestic use and normal exports and the input coefficient matrix for the production of processing exports, respectively. LCR policies in China 129 Table 5.7 Input-Output Table with Separate Production Accounts for Processing Trade   Intermediate use    Production Production Final use Gross for domestic of processing (C + I + output or use &normal exports G + E) imports exports DIM 1,2, ..., N 1,2, ..., N 1 1 1 Production for . domestic use .ZDD ZDP YE DP −XE−P & normal . Domestic exports (D) N intermediate 1 inputs . Processing . 0 0 EPEP exports (P) . N 1 . Intermediate inputs from .ZMD ZMP YMM imports . N Value added 1 VDVP Gross output 1 XE−PEP Source: Adopted from Koopman et al. (2012). The analytical solution of this extended IO model is g k b1 XEbPgIA bb DD ADP kgYE DP bk h \ =h \ h \ (8) hEP\g0IP g h hh \h gE\ h Its Leontief inverse can be computed as follows: s1 dIAss DD ADP gdLL DD DP g L=g h =gh (9) g f0IPD g hf gLL PP g h ds IAs () s1 DDDAPg IAsD g () 1 DD =h gh f0Ig IA IA DD D =s () s s1DDDP A I () d f g g g g h h s1 0 LIA A I DD DP DD DP PD PP =ss d f g g g g h h=d f g g g g h h s 0 1 130 Kun Cai and Zhi Wang Substituting Equation (9) into Equation (8), we have: XE IA YE AA E PDDDPD DDPp -=-- () +- -- () () 11 1 (10) Substituting Equation (10) into Equation (7), the total demand for imported intermediate inputs can be computed as MY AIAY EA AAEAE MMDDDD PMDDDDPP MP p -= -- () +- + -- () () 11 1 (11) MY AIAYEA AAEAE MMDDDDPM DDDDPP MP p -= -- () +- + -- () () 11 1 It has three components: the first term is total imported content in final domestic sales and normal exports, and the second and the third terms are indirect and direct imported content in processing exports, respectively. DVA (GDP) at the industry level can be computed as: DVA DVAAA IA IA A I AIA D P T V DV PDD DD DP V DDD v vs d vs d vs d t g g j l | | vd dd11 0 vvs d vs + d d t g g j l | | d d 1 1 AIAAA YE E V DDDDPV P T DP P (12) whereA v D and Av P are n × 1 vectors that denote the direct value-added coefficient vectors for domestic sale/normal exports and processing exports, respectively. Equation (12) can be rewritten as: DVAGDP AIAY EA AAEAE V DDDDP V DDDDPP V PP == -- () +- + -- () () 11 1 (13) DVAGDP AIAYEA AAEAE V DDDDP V DDDDPP V PP == -- () +- + -- () () 11 1 It also has three components: the first term is DVA in final domestic sale and normal exports, and the second and the third terms are indirect and direct DVA in processing exports, respectively. Notice that Y D = Y f + Y ef + Y ei; if we insert it into Equation (13) and rearrange, we can decompose each industry’s domestic value added, or GDP, into the following four parts: DVAGDP AIAYAIAYE AIAY V DDfV DDef Pf V DDei == -+ -- () +- - -- - () )( () 11 1EEA AAAE Pi V DDDDPV PP () +- + () - ()11 (14) LCR policies in China 131 where E Pf and E Pi are processing exports for final and intermediate products, respectively, and EE Pf Pi + = E P by definition. Equation (14) is an extension of the GDP decomposition Equation (5) with an additional fourth term that is generated from processing export production. 4 Data sources and estimation results To estimate the extended IO table that accounts for processing exports separately, we closely follow the Quadratic Programing procedures described in Section 2.3 of Koopman et al. (2012).13 The purpose of these procedures is to minimize a quadratic penalty function subject to a series of accounting identities and adding-up constraints based on official statistics. Standard national IOTs are used to determine sector-level production and trade, and information from trade statistics is used to determine the relative proportion of processing and normal exports within each sector; thus, all available data are used to split the national economy into processing and non-processing blocks, each with its own IO structure. After describing the data sources, we report the estimation results for China’s share of domestic content in its production and gross exports at the aggregate level, by firm ownership, by major destination countries and by manufacturing industries. 4.1 Data Inter-industry transaction and direct value-added data are from China’s 2007, 2012, and 2017 benchmark IOTs published by the NBS of China, while detailed trade data from 2007 to 2017 are from the General Customs Administration of China. The trade statistics are first aggregated from the 8-digit HS level to China’s IO industries. We partition both imports and exports in a given commodity classification into different parts based on the distinction between processing and normal trade in the trade statistics and on the UN BEC classification. A summary of such partitions as a percentage of China’s total exports and imports along with the share of processing exports during 2007–2017 is reported in Table 5.8 and Table 5.9. The UNBEC classifies each HS 6-digit product into one of three categories: intermediate inputs, capital goods and consumption products. In Table 5.8, we further decompose the first two categories into two subcategories: processing imports used for the production of processing exports that cannot be sold to domestic users by regulation and non-processing imports used for domestic sale and normal exports. Columns (1) to (5) in Table 5.8 sum to 100%; Columns (1) to (3) and (4) to (6) in Table 5.9 sum to 100%.14 These data are important parameters in our optimization model and the key information to understand our estimates of domestic and imported content shares in Chinese exports, especially their cross-sector heterogeneities and changing trends over time. 132 Kun Cai and Zhi Wang 4.2 Domestic contents in aggregate exports Table 5.10 presents the results for the decomposition of aggregate foreign and domestic value-added shares in 2007, 2012 and 2017. The estimated aggregate DVA share in China’s total gross exports was 64.2% in 2007, 65.2% in 2012 and 69.8% in 2017. Such numbers for merchandise exports Table 5.8 Major Imports Share Parameters Used in Domestic Content Estimation, 2007–2017 Year Imported intermediates % Imported capital goods % Imported Processing final exports as % For processing For normal For processing For normal consumption of total exports use exports use % exports (1) (2) (3) (4) (5) (6) 2007 37.8 47.3 8.3 4.0 2.6 51.6 2008 32.3 53.4 8.1 3.3 2.9 48.1 2009 30.8 53.9 9.8 2.1 3.4 49.8 2010 29.0 55.1 10.4 1.8 3.7 48.0 2011 25.6 58.4 10.3 1.6 4.1 45.2 2012 24.9 59.4 9.7 1.4 4.6 44.1 2013 24.9 59.7 9.6 0.9 4.9 41.6 2014 26.5 57.1 10.0 0.9 5.6 39.6 2015 27.1 55.6 9.9 0.9 6.5 36.8 2016 25.3 56.2 10.1 0.7 7.7 35.7 2017 23.6 58.5 9.6 0.7 7.6 35.1 Source: Authors’ calculations based on China Custom trade statistics and United Nations Broad Economic Categories (UNBEC) classification scheme. Table 5.9 Major Exports Share Parameters Used in Production Decomposition 2007–2017 Year Normal exports % Processing exports %  Share of Share of Share of Share of Share of Share of intermediates capital consumption intermediates capital consumption goods goods goods goods  (1) (2) (3) (4) (5) (6) 2007 49.92 13.39 36.69 33.28 34.26 32.46 2008 53.06 14.28 32.66 32.68 35.12 32.21 2009 46.80 15.23 37.97 31.20 36.34 32.46 2010 48.39 14.78 36.83 31.92 37.21 30.87 2011 49.38 15.01 35.61 31.82 37.34 30.84 2012 47.72 15.86 36.42 30.45 38.19 31.36 2013 46.95 15.86 37.19 31.75 36.57 31.67 2014 47.27 16.27 36.46 31.63 34.93 33.44 2015 46.87 16.87 36.26 32.77 36.37 30.86 2016 46.72 16.74 36.54 33.59 36.01 30.40 2017 47.86 16.59 35.54 33.85 35.86 30.29 Source: Authors’ calculations based on China Custom trade statistics and United Nations Broad Economic Categories (UNBEC) classification scheme. LCR policies in China 133 were 60.1%, 59.5% and 64.4%, respectively. For manufacturing products only, these estimated shares are lower from 59.2% in 2007 to 63.5% in 2017. In general, the estimated direct domestic value-added shares are less than one-third of the total domestic value-added shares. However, the difference between the direct foreign value-added share and the estimated indirect foreign value-added share was relatively small, indicating that most of the foreign content comes directly from imported foreign inputs and generates much less indirect value added compared to domestic value added. It is interesting that the DVA shares in normal and processing exports trend in opposite directions: the DVA share in China’s normal manufacturing exports increased from 82.9% in 2007 to 84.7% in 2017, but this share declined in processing exports from 36.6% in 2007 to 27.7% in 2017, which is completely different from the previous decades (more domestically produced inputs were used in China’s processing exports between 1997 and 2007, the DVA “share increased from 20.7% in 1997 to 37.0% in 2007, up by more than 16 percentage points” (Koopman et al., 2012)). This indicates that the increase in the DVA share between 2007 and 2017 in China’s total exports (approximately 5.3 percentage points increase) was mainly driven by the decline in processing exports in China’s total merchandise exports (decreased from 51.6% in 2007 to 35.1% in 2017) and Table 5.10 Domestic and Foreign Values Added: Processing vs Normal Exports (in Percent of Total Exports) Normal exports Processing exports Weighted sum 2007 2012 2017 2007 2012 2017 2007 2012 2017 Total exports (including service sectors) Total foreign value added 15.3 14.6 12.7 63.0 70.0 71.9 35.8 34.8 30.3 Direct foreign value added 4.9 4.7 4.8 58.0 66.4 69.5 27.7 27.2 24.0 Total domestic value added 84.7 85.4 87.3 37.0 30.1 28.1 64.2 65.2 69.8 Direct domestic value added 28.6 30.4 30.4 9.5 8.9 9.3 20.4 22.5 24.1 All merchandise Total foreign value added 16.7 16.6 14.9 63.0 70.0 72.0 39.9 40.5 35.6 Direct foreign value added 5.6 5.6 5.9 58.1 66.5 69.7 31.9 32.8 29.0 Total domestic value added 83.3 83.4 85.1 37.0 30.0 28.0 60.1 59.5 64.4 Direct domestic value added 23.4 22.1 22.3 9.4 8.9 9.2 16.4 16.2 17.6 Manufacturing goods (food processing sectors are excluded) Total foreign value added 17.1 17.1 15.3 63.4 70.3 72.3 40.8 41.4 36.5 Direct foreign value added 5.7 5.8 6.1 58.4 66.8 69.6 32.7 33.7 29.8 Total domestic value added 82.9 82.9 84.7 36.6 29.7 27.7 59.2 58.6 63.5 Direct domestic value added 22.5 21.4 21.6 9.4 8.9 9.2 15.8 15.7 17.0 Source: Authors’ estimates based on China’s 2007, 2012, 2017 benchmark input-output table published by the Bureau of National Statistics and Official China trade statistics from China Customs. 134 Kun Cai and Zhi Wang the increase in exports in services (increased from 14.4% in 2007 to 19% in 2017). The increase in the DVA share in China’s normal merchandise exports only played a relatively minor role (increased only approximately 1.1 percentage point from 84% in 2007 to 85.1% in 2017). This empirical finding may indicate that the various industrial policies and implicit LCR measures proposed in China during recent decades played no significant role in promoting DVA in China’s total exports, at least at the aggregate level during 2007–2017. 4.3 Domestic content in exports by firm ownership There is a significant change in export structure by firm ownership between 2007 and 2017. The share of private firms increased dramatically, from 21.3% in 2007 to 44.2% in 2017, more than doubling within 10 years. At the same time, the share of both state-owned and foreign-invested enterprises (SOEs and FIEs) declined from 18.9% to 9.1% and 56.3% to 45.3%, respectively. Both private firms and FIEs are the major players in China’s export success, and one may be interested in the DVA share in their exports. However, since there is no information on separate input-output coefficients by firm ownership, we cannot meaningfully distinguish foreign versus local firms within a sector. Instead, we provide an estimate of the DVA share of aggregate merchandise exports by firm ownership. By construction, the differences across firms of different ownerships are driven entirely by different degrees of their reliance on processing exports within a sector and the difference in the sector composition of their total exports (both are observed directly from the customs trade statistics). Estimates of the DVA shares by firm ownership for China’s merchandise exports are presented in Table 5.11. The results show that exports by wholly foreign-owned enterprises exhibit the lowest share of DVA, followed by Sino-foreign joint-venture companies (decreased from 44.2% and 56.7% in 2007 to 43% and 52.7% in 2017, respectively). Exports from Chinese private enterprises embodied the highest DVA shares (80.7% in 2007, 77.8% in 2012 and 81.1% in 2017), while those from state-owned firms were in the middle (approximately 70% in the three years). It is also interesting to observe that the variation of DVA share in normal exports is relatively small by different firm ownerships and over the three benchmark years (between 82% and 85%). The weights of processing exports are the decisive factor behind the difference in DVA share between private firms and FIEs (private firms only have approximately 10% of their exports as processing exports, while approximately two-thirds of exports from FIEs are processing exports). Note that these are estimations based on the currently available information; better estimates can be derived once information on I/O coefficients by firm ownership becomes available. Compared to Table 5.11 of Koopman et al. (2012), the most noticeable feature of this table is the relatively stable DVA share in exports produced LCR policies in China 135 by FIEs from 2007 to 2017 (it slightly declines). From 2002 to 2007, the DVA share increased by more than 10 percentage points. It seems that FIE exporters did not source more of their intermediate inputs within China after 2007. This finding provides further evidence that the implicit LCR measures we described in Section 2 did not affect most FIEs’ decision to source their production inputs outside China at the aggregate level. Their use of imported inputs increased during this period. 4.4 Domestic content in Chinese exports by trading partners By assuming that DVA shares within a given sector and export regime (normal/processing) are the same for all destination countries, we can further estimate the DVA share in China’s exports to each of its major trading partners. However, the variation by destination is driven solely by China’s export structure (share of processing exports and sector composition) to each of its trading partners. The results for China’s total merchandise exports to each of its major trading partners are reported in Table 5.12 in increasing order of the estimated weighted DVA share in 2017. Hong Kong, Singapore, Japan, the United States and Korea are the top five in both 2012 and 2017, with less than 60% of China’s DVA embodied in its imports from China. China’s exports to all emerging economies and developing countries embodied much higher DVA than its exports to OECD countries. The difference is more than 10 percentage points. Interestingly, the DVA share uniformly increased in China’s non-processing exports to all its major trading partners from 2012 to 2017, while it uniformly declined for processing exports to all countries. This information suggests that the LCR policies we discussed in Section 2 did not reduce exporting firms’ sources of raw materials, parts and components around the world, at least in China’s production of processing exports. 4.5 Domestic content in Chinese exports by industries To see if there are interesting patterns at the sector level, Table 5.13 reports, in ascending order of the weighted DVA share of 2012, the value-added decomposition in Chinese manufacturing exports by industry in 2012 and 2017, respectively, together with the shares of processing trade and foreign invested firms in each sector’s exports and the sector’s share in China’s total merchandise exports. Because the sector classifications are more consistent between 2012 and 2017 than those in 2007 due to a new version of industrial classification in China (CSIC, 2002 to CSIC, 201115), we present the results of each sector for the years 2012 and 2017 only. Thirteen out of the 68 manufacturing industries reported in Table 5.13 had a share of DVA in their exports of less than 51% in 2012, collectively accounting for 34% of China’s total exports. It is worth noting that more than half of these industries are relatively sophisticated and high-tech, such as 136 Kun Cai and Zhi Wang ts s total e of expor m ownership ts by fir in China’ expor 32.1 12.4 9.1 2.1 44.2 100.0 36.5 15.0 12.4 2.5 33.6 Shar 100.0 38.0 17.8 18.9 4.0 21.3 100.0 m Ownership (%), 2007, 2012 and 2017 otal T domestic value added 43.0 52.7 72.7 78.5 81.1 64.5 40.9 53.0 69.8 74.1 77.8 59.6 44.2 56.7 72.0 73.1 80.7 60.2 eighted sum ect Dir domestic eau of National Statistics and Official China W value added 12.5 15.1 20.6 20.1 21.3 17.6 11.4 14.3 19.2 19.8 20.8 16.2 13.5 15.4 19.9 19.1 22.2 16.5 otal T domestic value added 28.0 29.2 22.4 32.5 30.2 28.2 30.0 31.2 28.4 33.9 29.8 30.1 published by the Bur 36.1 38.5 39.4 41.8 42.2 37.1 ts by Fir Processing chandise Expor ect domestic value added table 9.0 9.6 9.2 9.4 9.5 9.2 8.7 8.8 10.3 9.4 9.6 8.9 Dir -output 11.4 8.8 9.8 9.5 10.3 10.0 T domestic otal value added otal Mer 84.8 84.8 84.2 83.5 85.6 85.2 82.6 82.5 81.7 82.3 84.3 83.4 83.8 83.6 83.4 83.1 84.9 83.4 processing alue Added in T oneect N Dir domestic value added 22.0 22.5 23.2 21.2 22.3 22.3 21.8 21.8 21.8 21.9 22.3 22.1 23.8 22.9 23.4 22.4 23.5 23.5 ts s 2007, 2012, 2017 benchmark input’ ts in e of Domestic V e of . Shar processing expor total expor 73.5 57.8 18.7 9.9 8.0 36.2 79.2 57.5 22.3 16.9 12.0 44.7 83.0 59.7 25.9 24.3 9.7 50.1 eign owned ms ms ms ms ms ms ms om China Customs ms eign owned ms Shar e fir e fir e fir eign owned ms ’ estimates based on China ms ms 5.11 Wholly for Joint ventur Collectively owned fir ms ableT Private fir All fir 2012 Wholly for Joint ventur Collectively owned fir Wholly for Joint ventur Collectively owned fir ms 2017 State owned fir State owned fir ms Authors Private fir All fir All fir : 2007 State owned fir Private fir ceSour trade statistics fr LCR policies in China 137 , 2012 and 2017 11.3 2.0 6.1 19.1 4.5 1.8 1.6 2.2 14.3 2.1 1.7 1.9 1.7 1.3 1.6 3.0 3.1 1.5 1.9 5.6 3.2 6.1 2.3 ts to 100.0 % of expor the world centages 2012 14.0 2.0 7.6 17.6 4.3 1.7 1.3 2.0 14.8 2.1 1.7 1.8 1.6 1.6 1.7 3.8 2.4 0.8 2.2 5.5 3.2 3.9 2.5 100.0 , as Per 58.1 59.0 59.7 63.7 63.7 74.0 76.6 74.8 75.1 76.3 64.6 tnersrading Par eighted sumW 2017 51.3 53.3 59.1 64.4 67.0 67.6 68.4 71.3 74.6 76.4 77.1 79.9 78.0 2012 41.9 50.6 56.6 54.5 54.7 56.6 58.7 59.4 60.0 64.5 64.1 68.4 66.3 68.4 69.6 70.3 71.2 71.5 73.2 74.2 75.2 75.5 76.0 59.6 eau of National Statistics and Official China ts to its Major T 2017 28.8 24.7 29.8 28.1 29.3 28.0 28.0 27.8 28.1 27.1 27.6 27.8 29.6 28.9 27.2 26.7 29.1 25.3 28.3 27.8 27.3 29.6 29.7 28.3 Processing 2012 28.8 27.4 32.3 29.9 31.8 32.7 30.1 29.6 29.9 31.8 31.3 32.8 33.7 32.6 28.2 29.7 32.1 31.6 30.1 30.4 30.8 33.5 33.4 30.2 published by the Bur chandise Expor table 86.8 -output on processing 2017 83.8 83.0 85.6 85.3 85.2 84.5 84.5 85.5 85.6 85.8 85.2 85.1 84.5 85.0 84.7 83.9 85.3 85.9 85.3 85.4 85.7 85.4 85.2 oss Mer N 2012 81.7 82.2 85.3 84.1 82.0 81.6 82.4 82.8 83.9 83.5 84.1 83.9 82.3 82.6 82.6 82.1 81.7 82.8 84.5 83.7 83.3 83.9 84.0 83.4 e in Chinese Gr % of processing 2017 59.2 51.0 50.4 46.3 46.9 44.6 36.8 36.7 36.8 31.8 31.3 29.3 24.9 19.0 14.6 17.0 16.9 14.8 18.8 14.2 9.3 16.7 13.3 36.3 ts alue Added Shar expor 2012 75.2 57.7 54.1 54.6 54.4 51.1 45.3 44.0 44.3 36.8 37.8 30.3 32.9 28.6 23.9 22.6 21.1 22.1 20.7 17.8 15.3 16.8 15.8 44.7 2017 s 2012 and 2017 benchmark input’’ estimates based on China . Domestic V th Africa ince ope/Central Asia om China Customs ov 5.12 e Authors ea Rep n Eur : able gion description ce ear -Saharan Africa Re Y Hong Kong Japan United States Mexico est EU15 T Singapor Taiwan pr Australia/NZ Rest of OECD Rest of Southeast Asia Thailand Brazil Indonesia RUS Middle East/Nor otal Kor East EU12 W Rest of Latin Amer/Caribbean India Philippines Sub Rest of South Asia Easter T Sour trade statistics fr 240 Lili Yan Ing and Rui Zhang Gopinath, G., Neiman, B., 2014. Trade adjustment and productivity in large crises. American Economic Review 104, 793–831. Grossman, G.M., 1981. The theory of domestic content protection and content preference. The Quarterly Journal of Economics 96, 583–603. Head, K., Mayer, T., Melitz, M., 2022. The Laffer curve for rules of origin. https:// scholar.harvard.edu/sites/scholar.harvard.edu/files/melitz/files/hmm_roo_laffer_ shared.pdf Ju, J., Krishna, K., 2005. Firm behaviour and market access in a free trade area with rules of origin. Canadian Journal of Economics/Revue Canadienne d’économique 38, 290–308. Krishna, K., Itoh, M., 1988. Content protection and oligopolistic interactions. The Review of Economic Studies 55, 107–125. Lahiri, S., Ono, Y., 1998. Foreign direct investment, local content requirement, and profit taxation. The Economic Journal 108, 444–457. Qiu, L.D., Tao, Z., 2001. Export, foreign direct investment, and local content requirement. Journal of Development Economics 66, 101–125. Yang, C., 2021. Rules of origin and auto-parts trade. https://chenying-yang.com/ RoO_autoparts.pdf A.1 Equilibria with and without LCR In the equilibrium without LCR, the domestic price index of composite input purchased by sector k from sector s,Q D ks ,, is given by PC z D ks s s i is ks s ss , ,.uj h h h h t y j j j j y y y y y y y y h jj h h h hh 1 1 1 1 b (A1) In contrast, the same price index in the equilibrium with LCR is given by: PC z C z D ks s s i is ks i is s s s , , f ff ff f f ff f f f f f f f ff ff f f f f 1 1 nn n NB sCNC kks D ks s s i is ss s k PC , , , f f f f f f f f j j k k k ] \ \ \ \ f ff ff f f f f 1 1 1 1 ff f f OG nzz kF PC ks D ks s s i ss s , , ,, f f f f f f f f j j k k k ] \ \ \ \ = ff ff f f f 1 1 1 1 ff f f OG n ii S ks i i S ks z C z s s s ff f f f f f f f f f fff f f f f f f f j j k k k ] \ \ \ \ * f ,, 11 1 1 ff n ff f f s kF, (A2) where ΩΩ NB s C s , and ΩNC s denote the sets of firms that find their LCR nonbinding, firms that decide to comply with their binding LCR, and firms that Appendix A Appendix: Describing the full equilibrium 242 Lili Yan Ing and Rui Zhang decide not to comply, respectively. We can characterize the compliance decision of firm i based on iS sii sii si s s d f g g gg ghgh ghgh ghg g gh g g g NB COG N if if and ,, ,, ff ff f f f f 1 1 1 1 CC OG if and sii si S s s ghgh ghg g gh gf g g f f g g g g h g g g g ,, , ff f f f f 1 1 1 1 (A3) where Sz zS ss kks ks s kks k s s    OG OG sOGOG ,,, ,, , ,, f ff ff f f f ff dff ff dff fg fg 1 1ss , and f f b b ks D ks ks PX s , ,, ff f ff f g b b 1 Input demands for goods produced by sector s depend on the sizes of other sectors k and the input-output linkage between k and s. Meanwhile, we assume that the final consumption expenditures XF,s are fixed. So the input market clearing condition is XC zP ks D ks in ii k k i nk D nk k kk ,, , , hn n g h h n m j jn g hn m j j n jj jhgg g gg t 11 yyy g g g g g g h \ \ \ n g k Xnk 1,, (A4) and XC z ks D ks in ii k k i nk k k ,, , b nn bb b n b b b g g u b b b b b g g u b b b ff gghgh h h hOG 111 1 1 b bbb n bb nh b b f f f f f g g g hb b f hh gghgh h k k k PX D nk nk D ks i ii k k ,, , hNC bb g g u b b b b b g g u b b b nh b b f f f f f g bb bbb hhh hh kk k C zPX i kD kk OG OG OG , ,, 11 gg g hb b b g g u b b b b b g g u b b b nm bb b f gghgh h h D ks i ii k k i k kk k C z , , hh NB C OG 111 1 b bbb nh b b f f f f f g g g hh k k PX D kk OG OG ,, (A5) Quantifying the impacts of local content requirements 243 Combining the conditions shown, we can define the two equilibria without and with the LCR. Proposition 1 (Equilibria without and with LCR). Given exogenous variables φi, zk,s, aDi k,, a Fi k,, bL,i, PF k , wk, β D ks ,, XF,k, and λi, and parameter θ and σ s, the equilibrium without LCR is a vector of price indexes { PD ks ,} that satisfies equations (A1) and (A4) for all k and s. The equilibrium with LCR is a vector of price indexes { PD ks , ′ } that satisfies equation (A2), (A3), and (A5) for all k and s. A.2 Equilibrium in relative changes We investigate the impacts of the LCR imposition by formulating the equilibrium in relative changes. The relative change of price index is: PmP D ks i i ks D s ii s s hh n n ,, g g gt t tyy u u u i o l l k j kk k m 11 1 1 1 1 u yy u ggh hNB ss s s s i i ks D s ii i i i k mP m by u u u i o l l b j kk n k j m m h h C NC , h n n gghn yy u 1 1 1 1 1 ,, ,, sD s ii Pk s h n n gghm yy u 1 1 1 1 1 kk k y u u u i o l lyOG PmP D ks i i ks D s ii s s hh j j ,, u g gg h hhh h j j j j g g gg g 11 1 1 1 g gg gg gg NB sNC U ghgh gh hh h j j j j gh ghh g g gg j g g 1 1 1 1 1 1 g gg g ggk s s s i i ks D s ii i mP k gC OG , ,, h j j FF PmP D ks i i ks D s ii s ss ,, . . ,, hh h hh hh hhhh h h h j j h h hh h h 11 1 1 1 h hh hh hh NB NC hh h hh h h h j j h hhh h h hh h h h h 1 1 1 1 1 1 1 h hh h hhhj s s s i i ks Dii i i mP hh C s , , . . ss mkF i ks * h hh h h h h hh hh hhh , ,, (A6) where mY X i ks i ks ks ,,, /= is the market share of firm i in sector k’s total input purchase from sector s, or the market share of firm i in the final consumption demand. The change in output prices of non-binding firms arises from the general equilibrium effect that affects the domestic composite input prices. For instance, non-binding firms reduce domestic input usages when the domestic 244 Lili Yan Ing and Rui Zhang composite input prices increase. In addition to the general equilibrium effect, the compliance cost penalties h i j directly inflate the output prices of the complying firms. Meanwhile, non-compliers are charged an ad valorem non-compliance fee of τ when selling to the upstream OG sector, which also increases their prices. According to the Cobb-Douglas formulation, the relative change in the cost of domestic composite input is PP DnD nD sn  ss gg g gg g g g,, . b (A7) Looking into the change in the local content λ  i of firm i and assuming that foreign composite input cost P F k is not affected by the LCR, we notice that h hh hh hh h h , . . i iD k i ii D k P P h h h hh h hhh hh h h h h g g g g h h hh h hh hh h h 1 1 1 1 1 1 11 1 1 h hh h h h h g g g ghh hh h h hh h h ii , (A8) which depends on PD k  given λi, γi, and θ. Hence, we can get vv v i i i '=  . With v i ' in hand, we can also calculate k i ': fg g ggg g gg g g i ii g gg g gfg \ ; ; [ \ \ \\g g g g \ ; ; [ \ \ \ \ \ \ \ \ \ \ ] ] ] 11 1 11 1 () . (A9) The change in domestic input share within firm i’s local content is m mm b b h m m i D k iD k i P P b b b ff f f b b ff f fgh () h h 1 1 1 . (A10) So we can rewrite the relative change in firm-level unit cost as C Pi P i D k ii kk iD k ii m m m , , , , b bn nn mnmn mnmn nv nvv vv vv 1 1 1 1 1 1 1 1 ,ifNC NB vv ,,ifC ik b b b n n m n nn v (A11) Quantifying the impacts of local content requirements 245 The total expenditure on domestic input in the LCR equilibrium can be written as: XYXC P ks ks D ks k k in ii i nk iD nk k kk ,, ,, ,    gg gh hh gg jj jj jgg jj 11 hOG 11 111 X CP XY nk i ii i kiD k s k kk     , ,, h j k k mg g hh gg j ngg jjj hNC OG OG OG,, ,,, , k i ii i kiD kk kk kk YCPXm b v b b gn hh gg j hh NB C OG OG OG gg jj    11 (A12) which helps to define X ks  ,given other variables. The following proposition describes the relative change of the equilibrium caused by the LCR imposition. Proposition 2 (Equilibrium in relative changes). Given endogenous variables Vv ii i ks ks YX ,, , ,, {} , exogenous variables {XF,s}, policy variables {τ, λi}, and parameters vb B ,,,s D ks {} , a relative change of the equilibrium caused by the LCR is a vector of price index changes PD ks  , that satisfies (A3), (A6), (A7), (A8), (A9), (A10), (A11), and (A12). Once we calibrate and obtain the values of {λi, γi, Y i ks ,, Xk,s}, {XF,s}, {τ, λi}, and {θ, σ s, β D ks ,}, we can evaluate the effects of imposing the MEMR LCR on firms and the economy. Index ADF Group Inc. v. United States of Brazil 2, 4, 22, 34–37, 63, 118, 174 America 162–163 Brazil, Russia, India, China, and South Agreement on Subsidies and Africa (BRICS) 174 Countervailing Measures (SCM BREXIT 4, 7 Agreement) 10, 63, 67, 150, Burkina Faso 54 153–154, 164 Ahafo Linkages Program 73 Canada 2, 6, 58–59, 60, 71, 72, 88, 91, Alta Ley National Mining Program 60 102, 112, 118, 158, 160–161, American Automobile Labeling Act 174; seealso North American (AALA) 91, 96, 102–103 Free Trade Agreement (NAFTA) Anglo American Corporation 73 Cargill, Inc. v. Mexico 159 Antofagasta Region Mining Cluster Chile 2, 6, 59–60, 61 (CMRA) 60 China: automobile industry 112, Archer Daniels Midland Company 116–119; domestic content in (ADM) v. Mexico 159–160 aggregate exports 132–134; Argentina 2, 5, 22, 30–31, 64, 174 domestic content in exports ASEAN–Australia–New Zealand by firm ownership 134–135; Free Trade Area (AANZFTA) domestic content in exports by Agreement 155 industries 135–142; domestic ASEAN–Japan FTA 165 content in exports by trading Association of Southeast Asian Nations partners 135; domestic content (ASEAN) 155 in exports/processing exports Australia 2, 6, 58–59, 60, 62, 66, 72, 174 128–131; domestic content in automobile industry: Argentina 22, production 126–131; impacts 30–31; China 112, 116–119; of LCR policies 8–9, 114–142; import substitution 2; Indonesia industrial policy initiatives 4, 4–5, 23, 37–39, 64; regional 115–126; integrated circuit parts use 91–97; rules of origin 4, industry 119–123; Made in 6, 7, 87–112; theoretical model China 2025 4, 8, 116, 123; of parts sourcing 97–111; United medical supplies 123–126; mining States 87–112, 119 industry 53, 61, 62, 63, 64, 68, Auto Pact 88 69, 71; pharmaceutical industry 123–126; telecommunications backward linkages 5, 10, 16, 49, 51, industry 123; trade agreements 54–60, 178, 212 10; use of LCRs 174 best efforts clauses 51, 54, 70 China – Audiovisual dispute 163 bilateral investment treaties (BITs) 10, 155 China – Auto Parts dispute 161 BMW 118–119 compliance 1, 7, 10, 49, 55, 57, 58, 90, Botswana 65, 69–70 97, 164 Index 247 compliance costs 11, 99–102, 111, 150, 152–153, 155–157, 164, 213–214, 229, 235, 244 165; seealso North American compliance decision 11, 101–104, 106, Free Trade Agreement (NAFTA) 212–216, 220, 225, 226, 228, Fundación Chile 60 230–233, 234, 238 comprehensive economic partnership General Agreement on Tariffs and Trade agreements (CEPAs) 10, (GATT) 9, 10, 63, 67, 150, 155–157, 164 152–153 Consejo Minero 60 General Agreement on Trade in Services Contemporary Amperex Technology (GATS) 9–10, 150, 153 Co., Limited 64 Germany 4, 18, 118 content protection see local content Ghana 6, 54, 55–56, 68, 73 requirements (LCRs) global value chains (GVCs) 54, 59, 60, Corn Products International, Inc. (CPI) 63, 71, 114 v. Mexico 159–160 government procurement 14, 20, 21, 22, 36, 39, 41, 115–116, 123, De Beers Diamond Consortium 65 126, 145, 146–148, 149, 152, de facto measures 159–160 154, 163, 178, 198, 215 demand-side policies 5, 50–51, 54–60, 66, 68, 72 Huawei 114 Democratic Republic of Congo 54 Huayou 64 Diamond Trading Company (DTC Hyundai 64 Botswana) 65 difference-in-differences (DID) estimator import licensing procedures 114–115 175, 208–211 incentives-based policies 5–6, 15, 18, 19, domestic employment requirements 49, 28, 51–53, 54, 58–60, 67, 71, 88, 51, 87 89, 97, 102, 115–116, 146, 147, domestic market obligations 15, 16, 17, 154, 164, 165, 178, 198, 212 18, 63, 214 India 4, 18, 118, 174 domestic processing requirements 63, Indigenous Land Use Agreements 59 146 Indonesia: automobile industry 4–5, 23, domestic value added (DVA) 8–9, 115, 37–39, 64; Benteng Program 9, 127–128, 130, 132–134, 145, 174; Deletion Program 9, 145–146, 164, 175 145, 174; effects of LCR policy on trade flows 174–198; energy energy sector 17, 81–86, 146, 147–148, sector 147–148; impacts of LCRs 153 on manufacturing firms/sectors equilibrium 241–245 212–238; impacts of LCRs on European Union and United Kingdom prices and welfare 235–237; Trade and Cooperation Act impacts of LCRs on sales, (TCA) 7, 112 value added, and employment European Union (EU) 63, 71, 90–91 233–235; Increased Use of export bans 53, 61–64, 146–147 Domestic Production Program 9, export licensing requirements 53, 63 145; LCR compliance decisions 220–238; LCR for upstream OG Finland 60 sector 215–220; lessons from Ford Motor Company 64 WTO and ISDS cases 161–163; foreign direct investment (FDI) 54 mining industry 61–64; Ministry forward linkages 6, 60–65 of Energy and Mineral Resources France 118 Regulation No. 15/2013 11; free trade 98 modern retail 149; National free trade agreements (FTAs) 7, 9–10, Car Program 9, 145, 174; 11, 12, 63, 67, 89, 112, 146, pharmaceutical industry 149; 248 Index telecommunications industry Mineral and Coal Mining Law 62 148; trade agreements 9–11, mining industry: Australia 58–59; 150–161, 164–165; trade and Botswana 65; Canada 58–59; industrial policies on LCRs Chile 59–60, 61; China 53, 61, 146–149, 164 62, 63, 68, 69, 71; economic Indonesia–Japan Economic Partnership impacts of LCRs 5–6, 68–71; Agreement 164–165 Ghana 6, 54, 55–56, 68, 73; Inflation Reduction Act 2 Indonesia 61–64; institutional integrated circuit (IC) industry 119–123 frameworks/coordination international investment agreements 66–68; policy implications of (IIAs) 146, 155, 158, 164 LCRs 71–74; South Africa interventions 18–20 23, 39–40, 56–57, 69, 73; investment treaties 10, 68, 155–157, 160 Tanzania 6, 54, 57–58, 64, 66, Investor-State Dispute Settlement 73; typology of LCRs affecting (ISDS) 10, 157–161, 165 50–53; Zambia 58 Ireland 91 Mining Skills Council (CCM) 60 Mobil Investments Canada Inc. and Japan 10, 60, 71, 118 Murphy Oil Corporation v. Canada 158, 159 Kazakhstan 67, 69 modern retail 149 Kenya 17, 54 Mozambique 54 labor 18, 51 Namibia 54 LG Energy Solution 64 national treatment obligation 67, local content requirements (LCRs): 160–161, 162, 164 alternative model 192–193; Newmont–International Finance China 114–142; definition 14, Corporation 73 50, 192–193; detrimental effects Nigeria 54 16–17; economic impacts 68–71; non-compliance fees 11, 213, 215–216, effects on exports 187–192; 220, 222, 223, 225–226, 230, effects on imports 182–186; 232, 233, 235, 236 effects on trade 174–198; North American Free Trade Agreement Indonesia 145–165, 174–198, (NAFTA) 6, 7, 87, 88, 89, 91, 212–238; literature review 97, 99–100, 102, 106, 108, 112, 176–178, 197; measuring 158, 159–160, 162, 214 impacts 53–54; mining industry 5–6, 23, 39–40, 48–74; operations 18 monitoring 35, 43, 58, 66, 73, Organisation for Economic Co-operation 149; policy implications 71–74; and Development (OECD) recent implementation 18–20, 21, 145 145, 174; scope/content 50; ownership requirements 68–69 theory/practice 1–3, 15–18, 145 performance requirements provisions localization barriers 14, 20 158–160 pharmaceutical industry 123, 149 Malaysia 62 prescriptive beneficiation requirements Mali 54 53, 54, 59, 69, 70, 72–73 medical supplies 5, 23, 39, 123–126 Production Development Corporation Merrill & Ring Forestry LP v. Canada 160 (CORFO) 59–60 METRO 4, 5, 21 PT Aneka Tambang 64 Mexico 2, 92, 102, 112, 118, 159; PT Industri Baterai Indonesia (IBC) 64 seealso North American Free public–private partnerships 51, Trade Agreement (NAFTA) 59–60, 70 Index 249 QMB New Energy Materials 64 trade agreements 1–2, 10, 145–165, 179–198 Regional Comprehensive Economic Trade-Related Investment Measures Partnership (RCEP) Agreement (TRIMS) 9, 10, 67–68, 146, 154, 155 150–151, 155, 164 regional content requirement (RCR) 87, Tsingshan 64 89, 90, 91, 97, 101, 103–106, 109, 111 United Kingdom 4, 7, 18, 91 regional investment agreements (RIAs) United States: automobile industry 4, 155–157 87–112, 118; mining industry Regional Trade Area (RTA) 88 60, 71; trade agreements 10; reporting requirements 55, 66, 73 trade war 114; use of LCRs requirement-based policies 51, 53–58, 174; seealso North American 60–65, 68 Free Trade Agreement resource allocation effects 17, 176, 214 (NAFTA) robustness check 193–197 United States Buy America Act 5, 16, rules of origin (RoOs) 1–2, 3, 4, 6, 7, 23, 40–41 87–91, 92, 95, 97, 98–99, 101, United States-Mexico-Canada Trade 104, 108, 177, 214 Agreement (USMCA) 1, 4, 6–7, Russia 118 87, 89–90, 102, 104–106, 109, 112, 214 Saudi Arabia 5, 23, 39 US – Renewable Energy dispute 161 S.D. Myers v. Canada 160 sourcing 16, 18, 21, 49, 51, 67, 69, 70, Valor Minero 60 73, 87–101, 112, 142, 149, 176, value-added activities 1, 3, 6, 15, 49, 198, 213, 214, 220, 221–223, 56, 58, 68–69, 70, 87, 116, 119, 226, 229, 232, 238 123, 135, 176, 212, 213, 215, South Africa 5, 6, 23, 39–40, 54, 56–57, 230, 233–234, 238 69, 73 value-added exports 128 South Korea 10, 64, 118 value-added shares 8–9, 132–134 Subsidies and Countervailing Measur es value-added tax (VAT) 22 (SCM) 10 Venezuela 64 supplier development programs (SDPs) Volkswagen Group China 64 51, 66, 74 supply-side policies 5–6, 50–51, 56, 66, West Java 66 68, 70, 72 World Trade Organization (WTO) 4, Sweden 60 8, 9, 10, 63, 67–68, 71, 81–86, 116, 146, 150, 154, 175 Tanzania 6, 54, 57–58, 64, 66, 73 telecommunications industry 4, 22, Zambia 6, 54, 58 34–37, 123, 148 Zimele Enterprise Program 73