Policy Research Working Paper 10937 Domestic Laws and Protectionism in Government Procurement Carlos Sebastian Barreto Cifuentes Katayoon Beshkardana Majed El-Bayya Lorenzo Rotunno Governance Global Practice October 2024 Policy Research Working Paper 10937 Abstract This paper examines how procurement rules affect inter- procurement policies characterized as protectionist nega- national trade, leveraging a novel dataset that characterizes tively correlates with trade openness across countries, in national laws across 141 countries. Text analysis of national both public and private markets. This protectionist effect laws on government procurement identifies prevalent is confirmed in gravity regressions. Countries with more protectionist measures such as preferential treatment for protectionist procurement laws are found to trade more domestic bidders and mandatory domestic sourcing. A domestically than from abroad in procurement markets. descriptive analysis reveals that 124 countries incorpo- Industry-level estimates suggest that these effects are stron- rate preferential treatment provisions, highlighting the ger for goods than for services. widespread nature of protectionism. The prevalence of This paper is a product of the Governance Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at lorenzo.rotunno@univ-amu.fr. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Domestic Laws and Protectionism in Government Procurement* Carlos Sebastian Barreto Cifuentes† Katayoon Beshkardana‡ Majed El-Bayyaꬷ Lorenzo Rotunnoꬸ JEL Classification Numbers: F14, H57, K49. Keywords: Government Procurement; Legal Analysis; Protectionism; Gravity. * We are indebted to Jose Signoret for his guidance and comments throughout the project. The project was implemented and the paper drafted while Lorenzo Rotunno was at Aix-Marseille University. The views expressed in this paper are those of the authors and do not necessarily represent the views of the IMF, its Executive Board, or IMF management. © 2024 The World Bank and International Monetary Fund. † Universidad Externado de Colombia. Email: carlos.barreto@uexternado.edu.co. ‡ Morgan State University, School of Business and Management. Email: katayoon.beshkardana@morgan.edu. ꬷ World Bank Group. Email: Mbayya@worldbank.org. ꬸ International Monetary Fund and Aix-Marseille University (on leave). Email: lrotunno@imf.org. I. Introduction Government procurement is a major market, accounting for 12% of world GDP in 2019 (Bosio et al. 2022). The COVID-19 pandemic and the ongoing Ukraine-Russian Federation crisis have further emphasized the role of governments in large sectors of the economy, including health, medical equipment and defense. 1 As a result, the use and allocation of government procurement funds have come under increased scrutiny due to their sheer size and importance. Governments often use public procurement to pursue socioeconomic objectives, such as supporting the development of small and medium-sized local firms or promoting environmentally friendly practices. However, these objectives can come at a cost to economic efficiency, and can result in protectionist policies that limit the participation of foreign suppliers in procurement auctions. These policies can take various forms, including inflating the price bid by foreign firms to provide a preference margin to domestic bidders, requiring products and services to have a certain level of local content, mandating the use of specific technologies, or imposing strict licensing requirements. The implementation of discriminatory regulations against foreign suppliers can result in significant barriers to trade and can increase the prices at which governments source goods and services. This raises important questions about the extent to which governments around the world implement domestic regulations that discriminate against foreign suppliers and the effects of these regulations on trade. To what extent do governments around the world implement domestic regulations that discriminate against foreign suppliers? And what are the effects of these regulations on trade openness and cross-border government procurement flows? This paper aims to explore these questions by examining rules that could be characterized as protectionist in government procurement and their impact on cross-border trade. By shedding light on these issues, we contribute to the debate on the role of protectionist policies in government procurement and the broader implications for international trade. To begin our analysis, we introduce a novel dataset about domestic laws that discriminate against foreign suppliers of goods and services in government procurement. For 140 1 According to reports from the U.S. GAO (Government Accountability Office 2021), the U.S. federal government spent $35 billion on medical supplies and pharmaceuticals in fiscal year 2020 alone to treat COVID-19 patients (half of the total increase in federal spending between 2019 and 2020). The Chinese government has spent over $21 billion on COVID-related expenditures between 2020 and 2022. The U.S. government has spent over $19 billion on defense under authorizations following the Ukraine-Russia crisis. 1 countries, we examined procurement laws that can come from national initiatives and ratifications of international treaties such as the Government Procurement Agreement of the World Trade Organization (WTO GPA) and Preferential Trade Agreements (PTAs). We read and coded the texts of these laws to identify instances of measures that can be characterized as protectionist. We classified them into three types commonly regarded as the main ones: lack of equal treatment provision, preferential treatment in favor of domestic bidders (e.g., preference margin), and domestic sourcing (e.g., “made in the country” provisions). A descriptive analysis of the data reveals that the majority of countries in our sample adopt protectionist policies in their procurement regulations. Of 141 countries, 124 have preferential treatment provisions, and 37 percent of the legal systems in the sample lack a basic national treatment clause, while 60 percent of the countries apply domestic sourcing provisions. There is significant overlapping between the domestic sourcing and preferential treatment provisions: of the 83 countries where the law includes a “made in the country” provision, 81 also provide for preferential treatment in favor of domestic bidders. We explore the main determinants of variation in protectionism across countries. As expected, we find that countries that participate in international agreements such as the WTO GPA and the EU have a lower protectionism index. That is also true of high-income countries (relative to the other, middle-income countries in the sample), and of countries with better institutions, as measured by the World Governance Indicators. The main objective of our empirical analysis is to estimate the relationship between protectionist laws and trade openness. To achieve this, we leverage the variation in the adoption of discriminatory laws across countries and identify the correlation with the import share in total expenditure. Data on the value of bilateral imports by industry are sourced from the ITPD-E database, which includes information on domestic sales. By considering the relationship between procurement laws and total trade (whose cross-border procurement component represents a small part), we aim to include the effects of laws on imports by domestic firms that win procurement contracts. In the appendix, we also report the results of the empirical analysis performed with data from the TiVA inter-country input-output (ICIO) tables, which permit to separate out estimates of cross-border government procurement flows from the rest of trade flows. Because of data limitations (such as the lack of time variation in the law data), as well as the lack of an empirical strategy to identify causal effects, we interpret our empirical results as suggestive of the association (or lack thereof) between protectionism in the law and trade flows. In the main cross-country regression specification, the outcome variable equals the value of imports from all countries, relative to total spending. We estimate the association between the import share in aggregate expenditure and indicators of protectionism in procurement law across countries, controlling for the influence of participation in the WTO GPA and EU, as well as other country-level variables, such as economic development and institutional quality. The results suggest that differences in the presence of protectionist laws correlate negatively with variation in trade openness across countries. The magnitudes of the implied relationship are substantial. We find that the import share in total expenditure is 9 percentage points lower in countries that have protectionist laws in government procurement (as classified by our three indicators) relative to a country with average trade openness. The results apply to most industries, even if they are stronger for services. When we estimate the association for 2 government and private purchases separately using the TiVA data, we confirm the finding that countries with more protectionist procurement laws import less in both public and private markets. To address some of the potential biases in our cross-country analysis, we adopt a second empirical approach based on a gravity equation for bilateral flows in government procurement, similar to Mulabdic and Rotunno (2022). This approach accounts for standard bilateral determinants of trade costs, such as distance, contiguity, and colonial history, as well as time-varying fixed effects specific to the exporters and to the importers, which absorb the influence of multilateral resistance terms and other country-specific determinants of trade (Head and Mayer 2014). We identify the impact of our country-level measures of protectionism on bilateral imports by interacting them with an indicator for domestic trade. Specifically, we examine whether protectionist measures lead governments, firms and households to buy relatively more from domestic suppliers than from foreign ones – an effect on the so-called “border effect” in gravity models (McCallum 1995). The gravity estimates confirm the suggestive evidence from the cross-country regressions. The interaction coefficient is positive indicating that the difference between domestic and international trade is larger in countries with protectionist procurement laws. This result is driven by the presence of preferential treatment provisions. For countries that give preferential treatment to domestic bidders, the estimates suggest that domestic purchases are around 16 times bigger on average than international ones, whereas for importers without a preferential treatment clause, domestic purchases are 7 times bigger than international ones. Industry-level results suggest that the negative relationship between trade and protectionism in government procurement affects industries differently. The correlation is more precisely estimated in the primary and manufacturing sectors than in the services. Using data from the international input-output tables from TiVA, we further find that the negative association between trade and protectionism in government procurement concerns public and private purchases in qualitatively similar ways. This finding corroborates our initial hypothesis that protectionism in government procurement can extend beyond the conduct of procurement auctions and affect the sourcing decisions of private firms as well (notably, the domestic ones who are awarded procurement contracts). The rest of the paper is organized as follows. Section II scopes out the existing literature on trade policy and government procurement. Section III describes the construction of the indicators of protectionism based on national laws on government procurement, and the trade data. We then present in Section IV the results from the cross-country regressions and from the estimations of the gravity model, both investigating the relationship between protectionism in the law and trade openness. Section V concludes by summarizing the evidence and outlining limitations and policy implications. II. Literature review This paper examines the importance of protectionist rules in government procurement across countries and assesses their impact on cross-border flows, drawing on insights from different 3 strands of the literatures in law and economics. An important body of research has sought to identify and quantify the effects of protectionism in procurement, both in terms of trade flows and efficiency. Theoretical papers in economics have investigated the impact of protectionism in government procurement on trade patterns, production, and welfare. The first general equilibrium model of trade with ‘distorted’ public expenditure, proposed by Baldwin (1970), suggested that protectionism in government procurement would have no effect on trade flows and specialization, as consumers offset the government's shift towards domestic goods by importing more. However, subsequent research has shown that the impact of protectionism in government procurement is more nuanced in settings with differentiated goods (Miyagiwa 1991) and incomplete information (McAfee and McMillan 1989). Furthermore, when increasing returns to scale are allowed, trade barriers in government procurement can change the patterns of specialization and alter the agglomeration of economic activity (Brülhart and Trionfetti 2001; 2004). Home-biased government procurement, which decreases demand for foreign varieties, can lower the price of imported goods and increase the size of government spending. However, this comes at the cost of welfare, as cheap foreign varieties are replaced by more expensive domestic ones (Larch and Lechthaler 2013). While the theoretical literature has largely assumed protectionism in government procurement, empirical work has attempted to identify the extent to which governments favor domestic firms over foreign ones in procurement contracts and how this behavior affects trade flows. We contribute to the literature that uses aggregated industry- and country-level data. Previous studies, such as those by Rickard and Kono (2014) and Gourdon and Messent (2019), have used aggregate trade and public spending data, without separating out cross-border procurement flows. These studies find that countries with higher government spending import less, suggesting the presence of protectionism in government procurement. Shingal (2011) finds evidence for home bias in procurement of services by Japan and Switzerland. We follow this strand of the literature by investigating the effects of variation in protectionist rules in government procurement across countries (irrespectively of whether this comes from participation in trade agreements) on trade openness. The literature also examines the effects of trade reforms aimed at liberalizing procurement markets, such as the WTO GPA and of PTAs, using a gravity model for bilateral trade. The evidence is mixed, with the WTO GPA having a liberalizing role (in Rickard and Kono (2014) and Gourdon and Messent (2019), but no significant effect found by Shingal (2011; 2015) for Switzerland and Japan), but little effect found for PTAs. Chen and Whalley (2016) show that even when controlling for the share of procurement value that is above the thresholds where the provisions of the agreement apply, the WTO GPA has a positive impact on trade. The estimates from these studies that employ gravity-like empirical specifications might be biased as they omit controls for price and country-level unobserved factors (multilateral resistance terms in the structural gravity model). Building on these shortcomings, Mulabdic and Rotunno (2022) employ a structural gravity model to estimate the effects of PTAs with provisions specific to government procurement, accounting for these omitted factors. Their findings confirm the previous evidence of a strong home-bias in government procurement, 4 5 5 relative to purchases by firms and households, and highlight significant positive effects of PTAs on cross-border procurement, especially in services. Moreover, they identify that policies that are “multilateral” promote cross-border procurement the most, and EU membership has the largest effect on trade in government procurement. 2 Given that our empirical analysis focuses on domestic laws that apply equally to foreign firms from different countries, we draw on a similar gravity structural model as Mulabdic and Rotunno (2022), but consider the effects on total trade (in government procurement and private purchases). A few studies have mobilized “micro” contract-level data to investigate the issue of home bias in government procurement. This work is normally on single countries or regions (e.g., the US, the EU, Norway and the Republic of Korea), and identifies cross-border procurement as contracts awarded to firms located abroad (or to a different region when home bias is defined locally). 3 Overall, the micro studies confirm the evidence from cross-country macro data. Using EU data and gravity-style estimation approach, Herz and Varela-Irimia (2020) find that government authorities award disproportionally more to firms located within their countries and their region. García-Santana and Santamaría (2021) confirm this result in a subsample of French and Spanish procurement contracts, and show that the public home-bias is driven by sub-national authorities.4 A related strand of the literature highlights the impact of winning government contracts on firms’ outcomes, thus stressing the economic relevance of government procurement. Lee (2021) and Ferraz, Finan, and Szerman (2015) find that winning a procurement contract increases long-term firm growth in employment, sales and value-added in Korea and Brazil, respectively. Furthermore, receiving procurement contract allows small and financially- constrained firms to acquire credit (di Giovanni et al. 2023), and increase investments (Hebous and Zimmermann 2021). The firm-level studies confirm how crucial the allocation of government contracts can be for productivity and economic efficiency. 2 One of the multilateral provisions with the strongest impact on cross-border procurement imposes the adoption of an e-procurement system. This evidence echoes the results of Lewis-Faupel et al. (2016), who find that regional governments with e-procurement systems in India and Indonesia are more likely to award contracts to firms from outside the region. 3 This standard definition of cross-border government procurement, which is consistent with the aggregated country-level approach, misses two other ways of exposure to foreign firms: (i) when the contract is awarded to the local subsidiary of a foreign firm; (ii) a foreign company indirectly participates in the execution of the contract by providing intermediate inputs to the domestic ‘winner’ (Cernat and Kutlina-Dimitrova 2016; 2020). Only (ii) could be measured in our “macro” approach using ICIO tables (see Riker (2013) for an application). 4 Taş et al. (2019) use contract-level data for members of the European Economic Area (EEA), Switzerland and Macedonia over the 2006-2016 to assess the role of WTO GPA. They find that the probability of awarding a contract to a foreign firm is significantly higher if the foreign firm is based in WTO GPA member country. They also provide suggestive evidence of a cost-reducing effect of the WTO GPA. 5 6 6 III. Data and descriptive evidence A. Protectionism in national law on government procurement To explore the relationship between trade openness and protectionism in domestic law, we compile a unique dataset that combines a text analysis of procurement laws from over 150 countries, with trade data from the ITPD-E database. The legal analysis involves the interpretation and coding of national procurement laws, with a focus on identifying provisions that discriminate against foreign firms, goods and services or give preferential treatment to domestic suppliers. With this approach we develop indicators that provide a comprehensive snapshot of protectionism in government procurement around the world as of 2021-2022. For our analysis, we focus on aspects that relate to discrimination and barriers to participation in procurement auctions against foreign firms, goods, and services. We identify three main ways the law can allow for discrimination against foreign firms, goods, and services in procurement auctions. For each area, a dummy variable is assigned with a value of one if the law is discriminatory and zero otherwise. The first feature of protectionism is the absence of a general provision for non discrimination between domestic and foreign firms (or goods and services) in government procurement. An indicator for preferential treatment further identifies laws that allow for discriminating against foreign bids, for instance by inflating them. This policy is equivalent to the application of an import tariffs on foreign goods (Cole, Davies, and Kaplan 2017). 5 In addition to preference margins, governments may favor domestic goods and services in procurement auctions by imposing domestic sourcing, for instance through “made in the country” provisions. To summarize the information from these indicators, we construct a composite indicator that equals one if the law is categorized as protectionist for all the three areas – lack of an equal treatment provision, preferential treatment and domestic sourcing. Table A.1 in the Appendix reports the indicators on the three policy areas (two instruments – preferential treatment and domestic sourcing – and a general provision) that governments often use to restrict access to procurement markets, as well as the values of the composite indicator. Protectionism is widespread in our sample of countries. Around 36 percent of the countries lack a general provision of equal treatment between domestic and foreign firms when it comes to government procurement. Only in 17 out of 141 countries where data is available, the law does not include preferential treatment clauses against foreign bidders. A domestic 5 The classic case of preferential treatment is the application of a preference margin. Argentinian law, for instance, says that government agencies must choose an offer from a national firm over that of a foreign firm if the price of the domestic bid is less than or equal to that of the foreign one increased by 15 percent. 6 sourcing provision is present in 58 percent of the sample. Because of the strong overlapping between preferential treatment and domestic sourcing (81 of the 83 countries that have domestic sourcing provision also have preferential treatment provisions in their procurement law), we use a composite indicator that equals one if the country is protectionist in all the three areas (discrimination, preferential treatment and domestic sourcing). Figures 1 and 2 display the relationship between the protectionist indicators derived from the coding of procurement law and relevant economic and institutional characteristics. The national legal provisions might reflect the commitments that governments have taken in international agreements concerning their procurement markets. This is supported by our data. Countries that are members of the WTO GPA, a plurilateral agreement promoting openness in government procurement, have a significantly lower level of protectionism. A portion of the difference between WTO GPA signatories and non-signatories can be attributed to EU member countries, as these represent 65 percent of WTO GPA members. As depicted in Figure 1, the average value of the protectionist index is indeed lower for EU countries. We also find that middle and low income countries have legal systems that are more protectionist than those of high income countries in our sample (as classified by the World Bank). In Figure 2, we explore the relationship between the protectionist indicators and institutional characteristics, as measured by the World Governance Indicators (WGI) database. We calculate the average percentile rank across the six governance indicators (voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption) for each country, using available data for each year. We then take the average across years for each country. Figure 2 shows a negative correlation between this overall measure of quality of institutions and protectionism in government procurement. Countries where procurement laws lack an equal treatment provision, provide for preferential treatment or impose domestic sourcing have worse institutions on average, with the difference being particularly large when considering domestic sourcing. Given the strong correlation between the different WGI indicators, we will use the average measure from Figure 2 in our empirical analysis. Overall, since the legal, economic and institutional characteristics considered in Figures 1 and 2 might well correlate with imports as well, we will control for their influence in our empirical analysis. 7 Figure 1. Protectionism in government procurement, participation in trade agreements, and income status Notes: The “average protectionist index” is the country-specific mean of the “protectionist” dummy equal to one if they country lacks an equal treatment provisions, and provides for preferential treatment and domestic sourcing. WTO GPA and EU memberships are as of 31st December 2019. The income classification comes from the World Development Indicators of the World Bank. The bars show the averages and the capped lines their 95% confidence intervals. Figure 2. Protectionism in government procurement and the quality of institutions across countries Notes: The protectionist indicators on the horizontal axis are defined in the paper. The “Ave. quality of institutions” index is constructed from the World Governance Indicators database as follows. For each year available in the database, we take the average of the percentile rank across the six governance indicators: voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption. These numbers are then averaged across years by country. The bars show the averages and the capped lines their 95% confidence intervals. 8 B. On the compatibility between the law data and the trade data and the interpretation of the empirical estimates Data on the value of bilateral trade by industry are sourced from the ITPD-E database. It collates only official and administrative data, and includes information on domestic trade. We consider annual data between 2000 and 2019, when trade in services is also measured. Considering this type of trade data that do not distinguish between different types of buyer (government vs others) has three main advantages over alternative approaches based on inter-country input-output (ICIO) tables that permit to split cross-border procurement from other trade flows. First, using aggregate trade data allows us to incorporate in our estimates the effect that protectionist laws in government procurement can have on the sourcing strategies of domestic firms. This is important because while governments might not vary much in their propensity to award contracts to domestic (rather than foreign) firms, they can impose different restrictions to the input sourcing of the awardees, for instance through “made in the country” provisions. Second, bilateral imports are computed from administrative data, while the trade data from ICIO tables are estimated based on different proportionality assumptions (Antràs and Chor 2022; Johnson 2018). Finally, data from ITPD-E database cover all countries in the world. We can thus include in the analysis all the countries for which we have coded information on procurement law (around 140 countries), whereas the ICIO data are generally available for a subset of countries. For these reasons, we carry out our baseline empirical analysis with the aggregate trade data, and discuss the results using estimated data for cross-border procurement flows from ICIO tables in the Appendix. Before moving on to the econometric analysis linking the variable on protectionist rules in government procurement and the variable on cross-border trade flows, it is important to address some caveats that could limit the significance of a relationship between the two variables. First, it should be noted that the data resulting from the analysis of the text of national laws only captures ‘de jure’ protectionism. As with many coding of legal texts, there may be discrepancies between the text of the law and its implementation. Our investigation explores whether the legal provisions have empirical bite in reducing trade openness. Furthermore, the text analysis only provides a static picture of the national law on government procurement as of 2021-22, whereas the annual trade data cover the 2000-2019 period. Although we know the year a decree was first promulgated, we have no information about possible amendments introduced or how the law looked before the decrees entered into force. Therefore, in our empirical analysis, we are assuming that the characteristics coded in 2021-22 apply over the entire period of the trade data. Any deviation from this assumption, which is not systematically related to the determinants of trade, could introduce measurement error and work against finding a significant association between our protectionism indicators and trade openness. Another partial discrepancy between the law and the trade data concerns the object of discrimination, which can refer to either firms or the goods and services being procured. In 9 10 10 the trade data, the country of origin corresponds to the place of origin of the good or service. 6 The indicators of protectionism can apply to the nationality of firms, goods and services, or both. For instance, the (lack of) national treatment indicator applies to both firms and goods/services, while provisions for preferential treatment apply to firms. Domestic sourcing obligations, on the other hand, refer to the nationality of goods and services and are consistent with the definition of nationality used in the trade data. The composite indicator includes rules that depend on the nationality of both firms and goods and services. By investigating also the impact of each type of protectionist measure, we can provide suggestive evidence on the role of the different nationality principles. Because of these data limitations, our empirical strategy relies on simple cross-country regressions and gravity models of bilateral trade. Our estimates are interpreted as correlations suggesting the existence (or lack thereof) of a relationship between trade openness and the protectionism measures. Reverse causality and omitted variables can introduce biases that prevent us from interpreting the estimates as causal. Countries whose markets have been historically open to foreign firms might have designed their law in ways that accommodate such a trend, thus creating a link going from trade openness to the (low) protectionism level of the law. Furthermore, unobservable country characteristics (for instance, legal aspects that are not measured by the coding of national laws) that correlate with both the coding of the law and the trade flows can influence the estimates of the relationship of interests. For these reasons, the estimates found in the econometric analysis are suggestive of the relationship between trade openness and protectionism of the law, without indicating causal effects. IV. Results on the relationship between trade openness and protectionism in national laws on government procurement We move beyond descriptive evidence and examine the relationship between protectionism in government procurement, as measured by a coding of the national law, and trade flows. We first estimate this relationship in cross-country regressions while controlling for potential confounding factors that could correlate with both protectionism and trade flows, as demonstrated in the descriptive analysis. In a second step of the econometric analysis, we estimate how the protectionist indicators relate to bilateral trade flows in a gravity model. 6 This is the case also for services sold by affiliates of multinational firms (what is commonly referred to as “commercial presence” or “Mode 3” of service trade). This means that, for instance, the purchase of business services by the Brazilian government from the Brazilian affiliate of a U.S. multinational would be recorded as domestic procurement. 10 A. Cross-country regressions We use the import share in total spending our dependent variable in cross-country regressions. Since the coded law should only distinguish between domestic and foreign firms, treating all imports equally irrespective of their source is an appropriate choice. The import share in government procurement is averaged over the 2000-2019 and taken in logs. Table 1 presents the results of the cross-country regressions of the import share on indicators measuring protectionism in the three main areas coded from the national law on government procurement (discrimination, preferential treatment and domestic sourcing). Consistent with our expectations, we find a negative association between import openness and the different indicators of protectionism in government procurement laws. Countries with more protectionist laws in government procurement tend to have lower foreign penetration in the total value of expenditure. This negative correlation emerges as we control for the effect of participating in the WTO GPA and EU trade agreements, as well as for other country characteristics (even-numbered columns in Table 1). The estimated coefficients on the protectionist indicators are economically significant. Using the specification with controls in column (8) as the baseline, the estimates imply that the import share of total spending is 20 percent lower in countries whose procurement laws are classified as “protectionist” (i.e., they do not provide for an equal treatment provision, they give preferential treatment to domestic firms and adopt a domestic sourcing provision) than in other countries. Relative to the average import share of 56 percent, this difference amounts to an 11 percentage points lower share. The results also suggest that preferential treatment is the type of protectionist law that has the most negative effect on the import share. We include several control variables in our cross-country regressions to account for factors that may be associated with both protectionism and import penetration in government procurement. One set of control variables measures a country's participation in international agreements that can shape its national laws on government procurement, including EU membership (arguably the most comprehensive regional trade agreement) and WTO GPA membership. Because of the cross-country empirical setup, we cannot properly control for membership in other preferential trade agreements – which is something that we can do in the gravity estimations. While EU and WTO GPA members have less protectionist laws in our data (see Figure 1), we find that this type of memberships correlates negatively with trade openness. We also control for the level of economic development (GDP per capita, in logs), dummies for income status and the average percentile rank across the six World Governance Indicators. The results indicate that high-income countries (the excluded category) have higher import openness, whereas GDP per capita and institutions do not correlate significantly with openness. 11 Table 1. Import share of total expenditure and protectionism in government procurement (cross-country) (1) (2) (3) (4) (5) (6) (7) (8) Dep. variable: Ln(Import share in Discrimination Pref. treatment Dom. sourcing Overall indicator total spending) Protectionism 0.031 -0.096 -0.185* -0.227** 0.030 -0.153** -0.039 -0.185* (0.092) (0.088) (0.102) (0.103) (0.084) (0.073) (0.108) (0.108) EU -0.151 -0.047 -0.124 -0.147 (0.145) (0.153) (0.153) (0.151) WTO GPA -0.507*** -0.551*** -0.546*** -0.524*** (0.142) (0.131) (0.141) (0.142) Ln(GDP cap) -0.070 -0.063 -0.060 -0.073 (0.061) (0.063) (0.061) (0.059) Middle income -0.235* -0.223 -0.212 -0.202 (0.141) (0.135) (0.134) (0.138) Low income -0.460* -0.405 -0.410 -0.422* (0.253) (0.252) (0.248) (0.249) Institutions WGI 0.001 0.001 0.001 0.002 (0.003) (0.003) (0.003) (0.003) Obs 139 138 138 137 139 138 139 138 R-sq 0.00 0.23 0.02 0.24 0.00 0.25 0.00 0.25 Notes: The dependent variable is the import share in total spending averaged by country over the 2000-2019 period, in logs. The “protectionism” dummy equals the variables reported in Table A-1. EU and WTO GPA variables are indicators for being EU and WTO GPA members respectively as of 31st December 2018. The Ln(GDP cap) variable is the country’s GDP per capita, in logs, averaged across the years. Middle income and Low income are indicators for middle income and low income countries, repsectively. Institutions WGI is the average percentile rank across the six WGI indicators. All regressions include a constant term. Standard errors robust to heteroskedasticity are in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. These results are indicative of a negative correlation between protectionist laws on government procurement and aggregate trade openness. In section A of the Appendix, we investigate how this correlation holds up for purchases by governments and those by private agents, measured from the ICIO tables of the TiVA database. We confirm a negative correlation that is not significantly different across the public and private sector. Overall, the relationship between protectionism and trade openness is weaker in the restricted sample of 66 countries of the TiVA data than in the full sample – a finding that is confirmed when we use the aggregate data from the ITPD-E database on the 66 countries. Under the assumption that the TiVA data can capture (at least partly) true differences in purchasing patterns between governments and the private sector, the lack of differences in the results between these two buyers suggest that our indicators of protectionism in government procurement identify restrictions that operate also on the sourcing decisions of firms. To gain further insights into the relationship between trade openness and protectionism in government procurement, we leverage the industry dimension of the ITPD-E database. There are 170 industries in the data: 118 in manufacturing, 35 in the primary sector (agriculture, forestry, fishing and mining), and 17 in the service sector. Because of missing data on internal trade especially in primary and services industries, we estimate the cross-country regressions in Table 1 with controls by sector, and retrieve the coefficient on the protectionism indicator of interest. The results are displayed in Figure 3. The import shares in the manufacturing and in the primary sectors are lower in countries with laws in government procurement that provide preferential treatment to domestic firms and contain provisions for domestic sourcing, whereas there is no significant difference with respect to the inclusion of a clause on national treatment. Overall, the coefficients are often imprecisely estimated, especially for the service sector. 12 Table A-4 in Appendix summarizes the results of the cross-country regressions by industry. Because of missing or incomplete data on internal trade, we lose 14 industries: 5 in the service sector and 9 in the primary one. The table shows the five industry with the largest (in absolute value) statistically significant coefficients associated to the protectionism dummy variables, and some summary statistics on the industry-specific coefficients. The correlation between import shares and the protectionism dummies is significant in a minority of industries – 16 percent when we use our composite protectionism indicator. All significant coefficients except one have a negative sign. The industries with the largest significant coefficient are both in the manufacturing and primary sectors, whereas the estimates for service industries are never significant. The estimates suggest that on average the negative correlation between import shares and protectionism in government procurement is slightly stronger in the primary than in the manufacturing sector, with important heterogeneity as shown by the number of manufacturing industries in the top-five list in Table A-4. Figure 3. Import share and protectionism in government procurement by sector, cross-country regressions Notes: The dot is the point estimate of the coefficient on the protectionist indicator used in the cross-country regressions. The bars are 95% confidence intervals. The dependent variable is within-country average of the import share in the primary, manufacturing or services sectors, as indicated, in logs. The specification includes the full set of controls and the constant term (even-numbered columns in Table 1). Standard errors are robust to heteroskedasticity. B. Results from gravity estimations We further investigate the relationship between trade openness and protectionism in government procurement in a canonical gravity model for bilateral trade, where the unit of observation is a country-pair consisting of an exporter and an importer. While our cross- 13 country regression framework is suitable to estimate the relationship between trade and our protectionism indicators, which are constant across exporting country, recent extensions of the gravity model allow us to estimate the trade effects of “unilateral” policies and economic characteristics that do not discriminate across exporters(Heid, Larch, and Yotov 2021; Sellner 2019; Beverelli et al. 2024; Piermartini and Yotov 2016). We estimate the following gravity model: , = exp�, + , + , � + , (1) The dependent variable , equals the value of imports by importer j coming from exporter i in year t, which goes from 2000 to 2019. The , and the , are importer-year and exporter- year fixed effects that control for the multilateral resistance terms that are common to different structural gravity models (see the reviews by Head and Mayer (2014); and Anderson (2011)), income and other country-specific determinants of bilateral imports. Therefore, the importer-year fixed effects absorb the level effect of trade costs variables specific to the importer that do not vary across exporters, such as the indicators of protectionism in government procurement. The term , collects bilateral variables that measure the level of trade frictions between the importer and the exporter, and is the vector of coefficients associated to each variable. In our baseline specification, the bilateral trade cost term is defined as: , ≡ 1 + 2 , + 1 � × � + 2 � × , � (2) The vector 1 collects the coefficients associated to the matrix of bilateral time-invariant determinants of trade costs: distance (in logs) and dummies for internal trade flows, for countries that share a border, an official language, the legal origin and colonial history. These variables are sourced from the CEPII Gravity Database (Conte, Cotterlaz, and Mayer 2022). The coefficient on the internal trade dummy measures the “border effect” (McCallum 1995; Anderson and van Wincoop 2003) – how much domestic purchases (and sales) are higher than trade with other countries, after controlling for bilateral determinants of trade and multilateral resistance terms. Following the insights of Heid, Larch, and Yotov (2021) and Beverelli et al. (2024), we identify the effect of our importer-specific measures of protectionism in government procurement (the protectionism variable) by interacting them with the dummy for internal trade. The coefficient 1 associated with this interaction term measures how much the border effect varies between more and less protectionist countries in government procurement. Importers whose law is coded as more protectionism should display thicker borders in government procurement, indicating a higher gap between domestic and foreign purchases. Conceptually, this empirical strategy is close to the cross- country regressions, where we compared import shares in total spending across countries 14 15 15 with varying degrees of protectionism in government procurement. 7 As in the cross-country regression, we control for other importer-specific variables that are likely to correlate with both protectionism in government procurement and trade openness: dummies for membership in the WTO, WTO GPA, EU, and GDP (in logs), GDP per capita (in logs) and the average index of quality of institutions (fixed over time). These controls collected in the term , are thus interacted with the internal trade dummy. Finally, we control also for the influence of time-varying determinants of bilateral imports that come from entry into trade agreements: WTO membership and PTAs. Because of our interest in indicators that code national laws in government procurement, we include dummies for membership of both the importer and the exporter in the WTO, WTO GPA, PTAs without provisions on government procurement, PTAs with provisions in government procurement (except the EU) and the EU. The data on WTO and EU memberships come from the CEPII Gravity Database, whereas information on WTO GPA membership comes from the WTO website. Data on PTAs membership and provisions on government procurement are sourced from the Deep Trade Agreements Database (Mattoo, Rocha, and Ruta 2020). Following the literature, the gravity models are estimated with the Poisson estimator (hence the exponential function in the gravity equation), which permits to include zero trade flows (around 30 percent of our sample) and accommodates heteroskedasticity in trade data (Silva and Tenreyro 2006). Following Egger and Tarlea (2015), we cluster the standard errors by importer, exporter and undirected country pair (the ij and ji pairs belong to the same cluster). Unlike the literature that deals with estimation of trade effects of trade agreements (Baier and Bergstrand 2007; Piermartini and Yotov 2016), our empirical specification lacks bilateral fixed effects, because our variables of interest –indicators of protectionism in government procurement laws – are time-invariant (their interaction with the internal trade dummy would be absorbed by the country-pair fixed effects). This limitation of our empirical setting prevents us from credibly identifying causal effects. Besides adding an exhaustive set of control variables as explained above, we try to lessen potential omitted variable bias by further controlling for interactions between year dummies and the internal trade dummy in additional specifications. As Bergstrand, Larch, and Yotov (2015) argue, controlling for trends in the border effect can net out important variation, which in our case might well correlate with the interaction between internal trade and protectionism in government procurement laws. Table 2 reports the results of the gravity estimations on the yearly sample, using the composite indicator for protectionism in government procurement laws – equal to one if the laws lack a national treatment provision (“discrimination”), they provide preferential 7 Mulabdic and Rotunno (2022) employ a similar empirical specification to estimate the impact on bilateral government procurement of provisions in PTAs that are likely to be multilateral (that apply equally to members and non-members). 15 treatment to domestic bidders, and establish domestic sourcing clauses. In the first two columns, we submit the data to two standard gravity specifications that control for importer- year and exporter-year fixed effects: in column (1) with time-invariant shifters of bilateral trade costs, and in column (2) with country-pair fixed effects that absorb that influence of time-invariant factors. The coefficients have the expected signs. The positive and large coefficient on the internal trade dummy indicates that trade is disproportionally domestic even when controlling for other determinants of bilateral trade. Distance depresses trade, while sharing a border, an official language, and (to a lesser extent) the origin of the legal system and having had a colonial relationship boost bilateral imports. The specification in column (2), by controlling for all time-invariant characteristics, attenuates the endogeneity bias that can affect especially the effects of time-varying policy variables (Head and Mayer 2014; Baier and Bergstrand 2007). Entering PTAs with provisions on government procurement has strong positive effects on trade. The estimates suggest that when two countries are members of a PTA with provisions on government procurement, bilateral imports go up by 32 percent. This result can be interpreted as a trade-creation effect of trade reforms in government procurement (through purchases of foreign goods and services by governments or by domestic firms that win procurement contracts) or other related trade policy areas included in PTAs (e.g., liberalization of trade in services, investments). They are consistent with those of Mulabdic and Rotunno (2022), who estimate different versions of the specification in column (2) using cross-border flows in government procurement as dependent variable (and on the 1995-2015 period). Columns (3) to (6) present our main specifications, which include an interaction between the same-country dummy and the protectionism indicator. The estimates suggest that the border effect is higher in importing countries that are classified as protectionist in their procurement laws, but the difference relative to other importers is not statistically significant. The size of the interaction effect is slightly higher when we control for other importer-specific variables interacted with the internal trade dummy (columns (4) and (6)). Also, controlling for trends in the border effect in columns (5) and (6) does not alter the estimates significantly. Table 2. Bilateral imports and protectionism in government procurement, gravity estimations (1) (2) (3) (4) (5) (6) Dep. variable: bilateral imports Baseline With time trends Internal trade 2.208*** 2.167*** 2.425*** (0.158) (0.153) (0.338) Log(distance) -0.555*** -0.566*** -0.435*** -0.567*** -0.424*** (0.092) (0.094) (0.087) (0.095) (0.088) Common border 0.379*** 0.374*** 0.454*** 0.376*** 0.453*** (0.108) (0.110) (0.119) (0.108) (0.117) Common language 0.390*** 0.407*** 0.229* 0.407*** 0.245** (0.116) (0.118) (0.125) (0.116) (0.120) Common legal origin 0.058 0.054 0.165*** 0.055 0.169*** (0.055) (0.057) (0.049) (0.057) (0.051) Colony 0.321* 0.309 0.346 0.306 0.355 (0.179) (0.190) (0.223) (0.192) (0.220) WTO w/out GPA 0.782*** -0.136 0.755*** 0.724*** 0.761*** 0.715*** (0.267) (0.111) (0.285) (0.241) (0.281) (0.245) GPA 0.661* 0.035 0.608 0.553** 0.601 0.545** (0.341) (0.132) (0.383) (0.265) (0.385) (0.267) PTA w/out procurement 0.573*** 0.086 0.547*** 0.961*** 0.558*** 0.993*** (0.177) (0.093) (0.178) (0.189) (0.181) (0.198) PTA with procurement (except EU) 0.147 0.275*** 0.095 0.191** 0.085 0.205** 16 (0.125) (0.080) (0.120) (0.079) (0.129) (0.087) EU 0.869*** 0.156** 0.854*** 1.001*** 0.852*** 1.025*** (0.220) (0.062) (0.220) (0.159) (0.219) (0.159) Internal trade x prot. Dummy 0.185 0.351 0.184 0.342 (0.255) (0.277) (0.262) (0.281) Internal trade x WTO w/out GPA 0.990*** 0.854*** (0.259) (0.265) Internal trade x GPA 1.293*** 1.234*** (0.395) (0.403) Internal trade x EU 0.208 0.189 (0.234) (0.234) Internal trade x Ln(GDP cap) -0.097 -0.221 (0.135) (0.146) Internal trade x Ln(GDP) -0.251*** -0.259*** (0.086) (0.088) Internal trade x Institutions WGI -0.014** -0.009 (0.007) (0.008) Obs 496570 496676 496570 490371 496570 490371 Notes: The dependent variable is the value of bilateral imports. Annual data from 2000 to 2018. All columns include importer-year and exporter-year fixed effects. Column (2) include country-pair (directed) fixed effects. Columns (5) and (6) include interactions between the internal trade dummy and year dummies. Standard errors three-way clustered by undirected country-pair, importer and exporter in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. In Table 3, we show the results of the gravity estimations using each of the three indicators of protectionism in government procurement separately (lack of national treatment clause – “discrimination” ; preferential treatment and domestic sourcing). We report the coefficient on the internal trade dummy, but omit for brevity the coefficients on all the other bilateral covariates (the results are similar to Table 2). Like for the cross-country regressions, we find that preferential treatment provisions have strong and significant effect on protectionism in trade data, as measured by the difference between internal and international trade. The estimates suggest that for importers whose laws have preferential treatment for domestic bidders in government procurement, domestic purchases are around 16 times bigger on average than international ones, whereas for importers without a preferential treatment clause domestic purchases are 7 times bigger than international ones. Lacking a national treatment provision or having a domestic sourcing provision increases the border effect, but not significantly so. 17 18 18 Table 1. Bilateral imports and indicators of protectionism in government procurement, gravity estimations (1) (2) (3) (4) (5) (6) Dep. variable: bilateral imports Discrimination Pref. treatment Dom. sourcing Internal trade 2.463*** 1.736*** 2.542*** (0.364) (0.497) (0.363) Internal trade x protectionism indicator 0.275 0.257 1.050** 1.068** 0.155 0.177 (0.323) (0.323) (0.423) (0.416) (0.182) (0.179) Internal trade x WTO w/out GPA 0.943*** 0.802*** 0.749*** 0.593** 0.835*** 0.686*** (0.258) (0.270) (0.232) (0.247) (0.239) (0.254) Internal trade x GPA 1.132*** 1.077** 1.024** 0.954** 1.181*** 1.116*** (0.419) (0.422) (0.402) (0.403) (0.400) (0.406) Internal trade x EU 0.258 0.231 0.134 0.114 0.133 0.118 (0.310) (0.305) (0.217) (0.224) (0.234) (0.236) Internal trade x Ln(GDP) -0.244*** -0.252*** -0.250*** -0.261*** -0.232*** -0.243*** (0.084) (0.086) (0.078) (0.079) (0.082) (0.084) Internal trade x Ln(GDP cap) -0.093 -0.221 -0.124 -0.263* -0.147 -0.281* (0.141) (0.155) (0.145) (0.154) (0.146) (0.157) Internal trade x Institutions WGI -0.014** -0.008 -0.012 -0.006 -0.012 -0.006 (0.007) (0.009) (0.007) (0.009) (0.008) (0.009) Obs 490371 490371 486594 486594 490371 490371 Notes: The dependent variable is the value of bilateral imports. Annual data from 2000 to 2018. All columns include importer-year and exporter-year fixed effects. Column (2) include country-pair (directed) fixed effects. Columns (5) and (6) include interactions between the internal trade dummy and year dummies. Standard errors three-way clustered by undirected country-pair, importer and exporter in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. In Figure 4, we display the coefficient on the interaction between the internal trade dummy and the protectionism indicator obtained by estimating our baseline gravity regression by aggregate sector. The specification includes all the bilateral as well as the importer-specific controls (column (4) in Table 2 and odd-numbered columns in Table 3). We confirm the finding from 3: the border effect is significantly higher for countries whose law on government procurement has a preferential treatment provision, and this holds in the primary, manufacturing and service sectors. This effect is strong and consistent in the manufacturing sector, whereas it is less precisely estimated in the other two sectors. In services, the estimates suggest that importers without a national treatment clause, with preferential treatment and a domestic sourcing provision (what we classified as “protectionist”) have a lower border effect. This counterintuitive result is however reversed for importers with preferential treatment provisions, and is not significant for importers that impose domestic sourcing. The results by industry summarized in Table A-6 in the Appendix are less conclusive than the results by sector, because of the incompleteness of the data at this level of disaggregation.8 They nonetheless confirm that the border effect is greater for importers with preferential treatment provisions in government procurement, and more precisely estimated for manufacturing industries. 8 The PPML estimates converged for 152 industries out of the 177 in the ITPD-E database. 18 Figure 4. Border effect and protectionism in government procurement by sector Notes: The dot is the point estimate of the coefficient on the interaction between the internal trade dummy and an indicator for a protectionist feature of the importer’s law in government procurement. The bars are 95% confidence intervals. The dependent variable is bilateral imports in the primary, manufacturing or services sectors, as indicated. The gravity specification includes the full set of bilateral controls and interactions between the internal trade dummy and other importer-specific variables (see column (4) in Table 3 and odd-numbered columns in Table 4). All regressions include importer-sector-year and exporter-sector-year fixed effects. Standard errors are three-way clustered by undirected country-pair, importer and exporter. We also estimate the gravity specification using the TiVA data and distinguishing between government procurement and private bilateral imports. The results are reported and discussed in section A.2 in the Appendix. They are consistent with the evidence from the cross-country regressions suggesting that the negative relationship between trade and protectionism in government procurement is present for public and private purchases. If anything, the enlarging effect of protectionism on the border effect is slightly stronger for bilateral trade in private markets than in government procurement. V. Concluding remarks In this paper, we introduce a novel dataset on the protectionist content of national laws across 141 countries and investigate how it relates to trade openness. The information on national laws concerning government procurement was organized into three areas representing the main policy tools that governments use to discriminate against foreign firms, goods and services (lack of a national treatment clause, preferential treatment for domestic firms, and domestic sourcing). For each of these areas and for the combination of the three, we constructed a dummy variable equal to one if the protectionist character emerges from the law. The data reveals that protectionism in government procurement is widespread. Around 36 percent of the countries lack a general provision of equal treatment between domestic and foreign firms when it comes to government procurement. Only in 17 out of 141 countries 19 where data is available, the law does not include preferential treatment clauses against foreign bidders. A domestic sourcing provision is present in 58 percent of the sample. We find that the share of countries with protectionist laws in government procurement is lower among countries that participate in major international agreements that liberalize procurement markets, the WTO GPA and the EU, and in high-income countries. We also find that this share tends to be lower in countries with institutions of better quality, as measured by the World Governance Indicators. To further validate the national law data and investigate the variation in trade openness across countries, we estimate the relationship between import share in total spending and the protectionism index. Because the national law data were collected during 2021-22 while the trade data cover the 2000-2018 period, and in absence of an exogenous shifter for the protectionism variables, we interpret the empirical estimates as suggestive of the correlation between trade and the protectionist character of national laws. The results indicate a systematically negative correlation between import shares and the protectionism in government procurement, which is robust to controlling for the influence of other country characteristics and different definition of government procurement in the trade data. Lower import shares in government procurement are especially observed in countries where the law provides for preferential treatment of domestic bidders. These results from cross-country regressions are confirmed in estimations of gravity models for bilateral imports. In particular, we find that the “border effect” (how much the country’s spending market is local relative to international, after controlling for standard determinants of bilateral trade) increases with protectionism in government procurement, with the effect being again stronger for preferential treatment provisions. The empirical exercise that we propose in this paper is subject to a number of limitations. In addition to the lack of a credible identification strategy that could support a causal interpretation of the estimates, there are measurement issues in protectionism in government procurement. These can explain why our estimates reveal correlations between trade openness and protectionism as observed in the law that are imprecisely estimated, especially for some components of the law. With these caveats in mind, our findings have important implications for government procurement policies. The relationship between protectionism in government procurement and overall trade underscores the significance of procurement laws for the sourcing decisions of domestic firms, namely those who win procurement contracts. It also highlights the possibility that the presence of this type of protectionism can well mask the presence of protectionism in other policy areas, unobserved to us, that matter for overall trade. 20 List of references Anderson, James E. 2011. “The Gravity Model.” Annual Review of Economics 3 (1): 133–60. Anderson, James E., and Eric van Wincoop. 2003. “Gravity with Gravitas: A Solution to the Border Puzzle.” The American Economic Review 93 (1): 170–92. Antràs, Pol, and Davin Chor. 2022. “Chapter 5 - Global Value Chains.” In Handbook of International Economics, edited by Gita Gopinath, Elhanan Helpman, and Kenneth Rogoff, 5:297–376. Handbook of International Economics: International Trade, Volume 5. Elsevier. Baier, Scott L., and Jeffrey H. Bergstrand. 2007. “Do Free Trade Agreements Actually Increase Members’ International Trade?” Journal of International Economics 71 (1): 72–95. Baldwin, Robert E. 1970. Nontariff Distortions of International Trade. Brookings Institution. Bergstrand, Jeffrey H., Mario Larch, and Yoto V. Yotov. 2015. “Economic Integration Agreements, Border Effects, and Distance Elasticities in the Gravity Equation.” European Economic Review 78 (August):307–27. Beverelli, Alexander Keck, Mario Larch, and Yoto Yotov. 2024. “Institutions, Trade, and Development: Identifying the Impact of Country-Specific Characteristics on International Trade.” Oxford Economic Papers 76 (2): 469–94. Bosio, Erica, Simeon Djankov, Edward Glaeser, and Andrei Shleifer. 2022. “Public Procurement in Law and Practice.” American Economic Review 112 (4): 1091–1117. Brülhart, Marius, and Federico Trionfetti. 2001. “Industrial Specialisation and Public Procurement: Theory and Empirical Evidence.” Journal of Economic Integration 16 (1): 106–27. ———. 2004. “Public Expenditure, International Specialisation and Agglomeration.” European Economic Review 48 (4): 851–81. Cernat, Lucian, and Zornitsa Kutlina-Dimitrova. 2016. “TTIP and Public Procurement: Going beyond the Tip of the Iceberg.” CEPS (blog). March 9, 2016. ———. 2020. “Public Procurement - How Open Is the European Union to US Firms and Beyond?” CEPS Papers, CEPS Papers, , March. Chen, Hejing, and John Whalley. 2016. “The WTO Government Procurement Agreement and Its Impacts on Trade.” In The Economies of China and India, 93–117. WORLD SCIENTIFIC. Cole, Matthew T., Ronald B. Davies, and Todd Kaplan. 2017. “Protection in Government Procurement Auctions.” Journal of International Economics 106 (May):134–42. Conte, Maddalena, Pierre Cotterlaz, and Thierry Mayer. 2022. “The CEPII Gravity Database.” Egger, Peter H., and Filip Tarlea. 2015. “Multi-Way Clustering Estimation of Standard Errors in Gravity Models.” Economics Letters 134 (September):144–47. European Commission. 2017. Measurement of Impact of Cross-Border Penetration in Public Procurement: Final Report. LU: Publications Office of the European Union. Ferraz, Claudio, Frederico Finan, and Dimitri Szerman. 2015. “Procuring Firm Growth: The Effects of Government Purchases on Firm Dynamics.” NBER Working Papers, NBER Working Papers, , May. Fronk, Jared. 2014. “Do Procurement Agreements Work?” Working Paper, mimeo. Giovanni, Julian di, Manuel García-Santana, Priit Jeenas, Enrique Moral-Benito, and Josep Pijoan-Mas. 2023. “Buy Big or Buy Small? Procurement Policies, Firms’ Financing, and the Macroeconomy.” FRB of New York Staff Report, no. 1006. 21 Gourdon, Julien, and James Messent. 2019. “How Government Procurement Measures Can Affect Trade.” Journal of World Trade 53:679. Government Accountability Office. 2021. “A Snapshot of Government-Wide Contracting for FY 2020 (Infographic).” WatchBlog (blog). June 22, 2021. Head, Keith, and Thierry Mayer. 2014. “Chapter 3 - Gravity Equations: Workhorse,Toolkit, and Cookbook.” In Handbook of International Economics, edited by Gita Gopinath, Elhanan Helpman, and Kenneth Rogoff, 4:131–95. Handbook of International Economics. Elsevier. Hebous, Shafik, and Tom Zimmermann. 2021. “Can Government Demand Stimulate Private Investment? Evidence from U.S. Federal Procurement.” Journal of Monetary Economics 118 (March):178–94. Heid, Benedikt, Mario Larch, and Yoto V. Yotov. 2021. “Estimating the Effects of Non- Discriminatory Trade Policies within Structural Gravity Models.” Canadian Journal of Economics/Revue Canadienne d’économique 54 (1): 376–409. Johnson, Robert C. 2018. “Measuring Global Value Chains.” Annual Review of Economics 10 (1): 207–36. Larch, Mario, and Wolfgang Lechthaler. 2013. “Buy National or Buy International? The Optimal Design of Government Spending in an Open Economy.” International Review of Economics & Finance 26 (C): 87–108. Lee, Munseob. 2021. “Government Purchases and Firm Growth.” SSRN Scholarly Paper. Rochester, NY. Lewis-Faupel, Sean, Yusuf Neggers, Benjamin A. Olken, and Rohini Pande. 2016. “Can Electronic Procurement Improve Infrastructure Provision? Evidence from Public Works in India and Indonesia.” American Economic Journal: Economic Policy 8 (3): 258–83. Mattoo, Aaditya, Nadia Rocha, and Michele Ruta. 2020. “Handbook of Deep Trade Agreements.” World Bank Publications - Books. McAfee, Randolph, and John McMillan. 1989. “Government Procurement and International Trade.” Journal of International Economics 26 (3–4): 291–308. McCallum, John. 1995. “National Borders Matter: Canada-U.S. Regional Trade Patterns.” The American Economic Review 85 (3): 615–23. Messerlin, Patrick. 2017. “How Open Are Public Procurement Markets?” In The Internationalization of Government Procurement Regulation, edited by Aris C. Georgopulos, Bernard Hoekman, and Petros C. Mavroidis. Oxford University Press. Miyagiwa, Kaz. 1991. “Oligopoly and Discriminatory Government Procurement Policy.” American Economic Review 81 (5): 1320–28. Mulabdic, Alen, and Lorenzo Rotunno. 2022. “Trade Barriers in Government Procurement.” European Economic Review 148 (September):104204. Piermartini, Roberta, and Yoto V. Yotov. 2016. “Estimating Trade Policy Effects with Structural Gravity.” CESifo Working Paper Series, CESifo Working Paper Series. Rickard, Stephanie J., and Daniel Y. Kono. 2014. “Think Globally, Buy Locally: International Agreements and Government Procurement.” The Review of International Organizations 9 (3): 333–52. Riker, David. 2013. “Government Consumption and Global Value Chains.” Sellner, Richard. 2019. “Non-Discriminatory Trade Policies in Panel Structural Gravity Models: Evidence from Monte Carlo Simulations.” Review of International Economics 27 (3): 854–87. 22 Shingal, Anirudh. 2011. “Services Procurement under the WTO’s Agreement on Government Procurement: Whither Market Access?” World Trade Review 10 (4): 527–49. ———. 2015. “Econometric Analyses of Home Bias in Government Procurement.” Review of International Economics 23 (1): 188–219. Silva, J. M. C. Santos, and Silvana Tenreyro. 2006. “The Log of Gravity.” The Review of Economics and Statistics 88 (4): 641–58. Taş, Bedri Kamil Onur, Kamala Dawar, Peter Holmes, and Sübidey Togan. 2019. “Does the WTO Government Procurement Agreement Deliver What It Promises?” World Trade Review 18 (4): 609–34. 23 Appendix A. Empirical analysis on cross-border flows in government procurement estimated from ICIO tables The empirical analysis in the main text investigates the relationship between protectionism in government procurement and trade openness as measured in official statistics, which do not distinguish between cross-border government procurement and imports by firms and households. Our estimates thus incorporate also the effects of protectionist laws in government procurement on imports by the private sector, notably domestic firms that win procurement contracts but can be constrained in their input sourcing strategies, for instance by “made in the country” provisions. An alternative approach consists in mobilizing inter- country input-output tables (ICIO), for instance from the TiVA OECD database, to split cross- border imports in government procurement and imports by firms and households. Previous work on the measurement of cross-border government procurement at the aggregate level (Messerlin 2017; Mulabdic and Rotunno 2022) has also relied on ICIO tables. This approach has two main limitations in our setting. First, data from ICIO tables are estimated (while data on trade flows are from official statistics), thus generating estimation errors that could bias our results in directions that are difficult to predict. National input- output tables provide information on imports of goods and services (classified by industry) by purchasing industry and country, aggregated across sources. ICIO tables of the TiVA database distribute these imports across source countries through a ‘modified’ proportionality rule that distinguish between imports meant for intermediate use and imports for final use at the product level (Antràs and Chor 2022; Johnson 2018). Second, the TiVA database covers only 66 countries, thus reducing our sample of countries and hence the cross-country variation that we exploit in our estimates. Because of these limitations relative to the official trade data that we use in the main text (sourced from the ITPD-E database), we present the results using the ICIO TiVA data in the Appendix. A.1 Descriptive trends The TiVA OECD database provide information on country- and industry-level trade in goods and services by type of buyer and are available annually from 1995 to 2018. We use these data to construct measures of domestic and foreign flows in government procurement, by country of origin. Government procurement involves different types of purchases, including intermediate consumption, social transfers in kind via market producers of goods and services, and gross fixed capital formation. To define the scope of our empirical analysis, we follow Mulabdic and Rotunno (2022) and the European Commission (2017) and measure the intermediate consumption components of government procurement as the sum of purchases recorded under the “Public Administration”, “Health” and “Education” columns of the ICIO tables. This choice reflects the reality of many countries where government authorities and services are heavily involved in the provision of health and education services. Social transfers in kind correspond in the data to the final demand purchases recorded under the 24 25 25 “Government expenditures” column. From these, we remove purchases made from the “Public Administration, defense and social security” industry, which includes defense spending that is typically not subject to competition due to geopolitical and strategic reasons (Fronk 2014; Rickard and Kono 2014). 9 Finally, we acknowledge that our measure does not capture procurement that constitutes gross fixed capital formation, such as public investments, due to the lack of data in the ICIO tables. As part of our analysis, we compare cross-border trade in government procurement with that in private markets. For our purposes, private markets are defined as purchases made by all industries except for "Public Administration," "Health," and "Education," as well as purchases made by households and private non-governmental institutions for final demand. 10 As a first check on the suitability of the ICIO data for measuring government procurement flows, we calculate the standard measure of the value of government procurement relative to a country GDP. On average, government procurement accounts for 16 percent of a country’s GDP in our sample, which is consistent with previous estimates in the literature (Bosio et al. 2022). This suggests that our data is appropriate at least for assessing the size of government procurement. Figure A-1 displays the main measure of trade openness in government procurement: imports relative to the total value of government purchases. On average, approximately 13 percent of the total value of government procurement in a given year is spent on imported goods and services. This import penetration ratio has been increasing over time on average and varies significantly across countries. In some countries, cross-border procurement can account for as little as 5 percent of total government purchases, while in others, it can account for as much as 27 percent. 9 We include in our measure services that the “Public Administration” buys from itself for intermediate use. 10 We thus exclude gross fixed private capital formation (i.e., private investments) and changes to inventory for comparison with the definition of government procurement. 25 Figure A-1. Import share in government procurement Notes: Government procurement flows are measured as the sum of purchases recorded under the “Health”, “Education” and “Public Administration”, and the “Final government expenditure” column of the ICIO TiVA tables. Imports are purchases on goods and services coming from another country. To benchmark our findings on government procurement, we compare the level of trade openness in procurement markets to that observed in private markets. We define the import share ratio as the ratio of the import share in government procurement to the import share in purchases by private firms and households. Figure A-2 displays the trend in this ratio over time, and it shows that on average, the import share in government procurement is 60 percent of that in private markets. This confirms the existence of “home bias” in government procurement, which is in line with both empirical and anecdotal evidence. The average import share ratio across countries is relatively stable over time, but we observe significant variation across countries. While most countries import relatively less through their governments than through private firms and households, there are a few cases where the import share in government procurement is higher than in private markets (the import share ratio being greater than 1). These exceptions are mainly found in small developing countries in our sample, such as Brunei, Kazakhstan, and Cambodia. 26 Figure A-2. Import share ratio Notes: Government procurement flows are measured as the sum of purchases recorded under the “Health”, “Education” and “Public Administration”, and the “Final government expenditure” column of the ICIO TiVA tables. Purchases by private firms and households are the sum of values recorded under the other purchasing industries columns, and the “Final expenditure by household” column of the ICIO TiVA tables. Imports are purchases on goods and services coming from another country. So far, our country-level analysis using ICIO data has ignored heterogeneity across sectors. However, work by Mulabdic and Rotunno (2022) using TiVA data finds that government procurement is heavily concentrated in services, and that governments are particularly more home biased in their purchases of services than private firms and households. We confirm these patterns using our definition of government procurement, which excludes government final expenditures from the public administration and defense industry as part of these flows represent non-contestable procurement. The services share of government procurement is approximately 85 percent on average in our sample, while the same share in private markets is 60 percent. Figure A-3 shows the trend in the import share ratio in the services and goods (manufacturing and primary products) sectors. Relative to the total size of purchases in government and private markets, governments import 60 percent fewer services than private firms and households (relative to the total size of purchases in government and private markets). In contrast, governments appear to be less home biased than firms and households when it comes to goods, with the average import share ratios being above 1. However, this finding is reversed in a substantial 40 percent of our sample. 27 28 28 Figure A-3. Import share ratios by sector Notes: Government procurement flows are measured as the sum of purchases recorded under the “Health”, “Education” and “Public Administration”, and the “Final government expenditure” column of the ICIO TiVA tables. Purchases by private firms and households are the sum of values recorded under the other purchasing industries columns, and the “Final expenditure by household” column of the ICIO TiVA tables. Imports are purchases on goods and services coming from another country. “Goods” industries are those classified under chapters 01 to 33 according to the ISIC Rev.4 classification. “Services” industries are the remaining industries classified under chapters 35 to 98. Dots are averages across countries. The lower end of the bar gives the 1st decile, while the upper end gives the 9th decile of the distribution, for each year. In Figure A-4, we examine further the heterogeneity in the import share ratio across the 45 industries of the TiVA database, split into goods on the left-hand side, and services on the right-hand side of the Figure. For each industry and year, we estimate the mean import share ratio across countries and the associated confidence interval. The graph reports these for the initial and final years of the sample, without the fishing and mining industries because the average import share ratio for these is not significantly different from zero. 11 Two key findings emerge from these estimates. First, we find that import penetration in government procurement is generally lower than in private markets, with this being true for most of the goods industries as well. Specifically, in 1995, the cross-country mean import share ratio is statistically greater than 1 only in three goods and one services industries, and in 2018, this never happens. In contrast, the mean import share ratio is significantly lower than 1, indicating home bias in government procurement, in four goods and ten services industries in 1995, and in eight goods and eleven services industries in 2018. Overall, the import share ratio appears to decrease between 1995 and 2018, which is less visible at the 11 Statistical significance here and elsewhere in the text means that the estimate is different from zero at 5 percent significance level. 28 29 29 aggregate, possibly because of composition effects – that is, governments buying relatively more goods and services where its purchases are less home biased. Second, we find that the propensity to import differ significantly across public and private buyers in most industries – with 66 percent of the industries showing significant differences in 1995, and 58 percent in 2018. The most extreme cases of these differences are in the health and education service industries, where governments appear to source only 2 percent of what firms and households do (relative to the total size of purchases) from abroad. The analysis of import shares across industries reveals marked differences in import penetration between public and private buyers, suggesting the existence of home bias in government procurement. This finding is consistent with previous studies and confirms that governments import less compared to private firms and households, particularly in the service sector. 12 The econometric analysis in the ensuing section aims to assess whether the variation across countries in import exposure in government procurement is systematically explained by our measures of protectionism in national laws on government procurement. Figure A-4. Import share ratios by industry 1995 2018 12 This difference in import shares also reassures against the possibility of statistical indistinguishability between import values in government procurement and in private markets, which could be the result of the proportionality assumptions used in ICIO tables. 29 Notes: Government procurement flows are measured as the sum of purchases recorded under the “Health”, “Education” and “Public Administration”, and the “Final government expenditure” column of the ICIO TiVA tables. Purchases by private firms and households are the sum of values recorded under the other purchasing industries columns, and the “Final expenditure by household” column of the ICIO TiVA tables. Imports are purchases on goods and services coming from another country. The dots are estimated means by year and industry (the variation is thus across countries), and the bars represent the associated 95 percent confidence intervals. A.2 Econometric results As in the main text, our econometric analysis using the TiVA data consists in two parts. We first estimate cross-country regressions of import share in total spending (splitting the government and private sectors as buyers), and then estimate gravity regressions on bilateral trade flows. Before discussing the results for government procurement and trade in the private sector, we check the comparability between the TiVA data and the ITPD-E data that we use in the main analysis. Table A-1 reports the estimates of the baseline cross-country regression equation, with the full list of controls, for each of the protectionist indicators. Compared to our baseline results in Table 1, the negative correlation between import share and protectionism is much weaker in the smaller TiVA sample (around half of the baseline one). Importantly, in this restricted sample we lose all low income countries and part of the middle income ones. Using the ITPD-E (fist four columns) or the TiVA (last four columns) data does not change significantly the point estimates, thus suggesting that in the aggregate the two sources of data point to similar conclusions. 30 Table A-1. Cross-country regression of import openness on protectionism using ITPD-E and TiVA data (1) (2) (3) (4) (5) (6) (7) (8) ITPD-E data TiVA data Dep. variable: Pref. Dom. Overall Pref. Dom. Overall Ln(Import Discrimin. Discrimin. treatment sourcing indicator treatment sourcing indicator share in total spending) Protectionism -0.107 -0.280 0.003 -0.045 -0.229* -0.146 0.046 -0.172 (0.174) (0.294) (0.161) (0.220) (0.130) (0.183) (0.131) (0.159) EU -0.093 -0.009 -0.063 -0.072 0.133 0.224 0.198 0.164 (0.200) (0.203) (0.213) (0.207) (0.150) (0.160) (0.166) (0.163) WTO GPA -0.445* -0.499** -0.501** -0.495** 0.279 0.160 0.158 0.184 (0.245) (0.220) (0.249) (0.245) (0.195) (0.190) (0.205) (0.200) Ln(GDP cap) -0.240* -0.230 -0.230 -0.232 -0.252** -0.230* -0.237* -0.238* (0.138) (0.143) (0.151) (0.139) (0.123) (0.128) (0.127) (0.123) Middle -0.446* -0.453** -0.481** -0.471** -0.135 -0.194 -0.223 -0.171 income (0.238) (0.217) (0.227) (0.231) (0.218) (0.208) (0.202) (0.214) Institutions 0.004 0.004 0.005 0.004 0.004 0.005 0.006 0.005 WGI (0.007) (0.007) (0.007) (0.007) (0.006) (0.006) (0.006) (0.006) N 66 66 66 66 66 66 66 66 R-sq 0.20 0.21 0.19 0.19 0.23 0.20 0.20 0.21 Notes: The dependent variable is the import share in total spending averaged by country over the 2000-2019 period, in logs. In columns (1) to (4), the variable is constructed from the ITPD-E database. In the other columns, the variable equals the sum of “public” and “private” imports and total spending constructed from the TiVA data. The “protectionism” dummy equals the variables reported in Table A-1. EU and WTO GPA variables are indicators for being EU and WTO GPA members respectively as of 31st December 2018. The Ln(GDP cap) variable is the country’s GDP per capita, in logs, averaged across the years. Middle income is an indicator for middle income countries. Institutions WGI is the average percentile rank across the six WGI indicators. All regressions include a constant term. Standard errors robust to heteroskedasticity are in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. Table A-2 shows the estimates of the cross-country regression with the import shares in total government procurement (first four columns) and total private spending (last four columns) as dependent variable. As for the aggregate import openness (Table A-1), the correlation with protectionism in government procurement is overall negative but imprecisely estimated. Importantly, the coefficients have the same sign and are of similar magnitudes across public and private spending, suggesting that countries with more protectionist laws in government procurement imports less to similar extents in the public and private sectors. Assuming that the ICIO TiVA data can measure meaningful differences across the two sectors (as Figure A-4 seems to suggest), these findings indicate that the protectionist character of national laws as measured by our indicators constrains domestic private firms (including those that win procurement contracts) in their sourcing decisions. 31 Table A-2. Import share in government and private spending and protectionism in government procurement (1) (2) (3) (4) (5) (6) (7) (8) Buyer: Government Private Dep. variable: Pref. Dom. Overall Pref. Dom. Overall Ln(Import Discrimin. Discrimin. treatment sourcing indicator treatment sourcing indicator share in total spending) Protectionism -0.234 -0.286 0.111 -0.146 -0.234* -0.123 0.038 -0.176 (0.157) (0.276) (0.151) (0.186) (0.129) (0.175) (0.128) (0.159) EU 0.068 0.188 0.139 0.106 0.139 0.227 0.206 0.171 (0.186) (0.185) (0.193) (0.200) (0.147) (0.160) (0.164) (0.159) WTO GPA 0.160 0.040 0.034 0.059 0.286 0.165 0.163 0.189 (0.291) (0.241) (0.292) (0.285) (0.193) (0.194) (0.205) (0.199) Ln(GDP cap) -0.271* -0.248* -0.265* -0.255* -0.238* -0.215* -0.221* -0.224* (0.145) (0.148) (0.155) (0.145) (0.122) (0.127) (0.126) (0.122) Middle -0.534* -0.580** -0.643** -0.576* -0.117 -0.179 -0.203 -0.153 income (0.304) (0.257) (0.281) (0.292) (0.220) (0.213) (0.206) (0.216) Institutions -0.006 -0.006 -0.004 -0.006 0.005 0.005 0.006 0.005 WGI (0.007) (0.007) (0.008) (0.007) (0.006) (0.006) (0.006) (0.006) N 66 66 66 66 66 66 66 66 R-sq 0.22 0.21 0.20 0.20 0.25 0.21 0.21 0.22 Notes: The dependent variable is the import share in total government procurement (columns (1) to (4)) and and the import share in total private spending (columns (5) to (9)). Government procurement and the private sector are defined in the text. The “protectionism” dummy equals the variables reported in Table A-1. EU and WTO GPA variables are indicators for being EU and WTO GPA members respectively as of 31st December 2018. The Ln(GDP cap) variable is the country’s GDP per capita, in logs, averaged across the years. Middle income is an indicator for middle income countries. Institutions WGI is the average percentile rank across the six WGI indicators. All regressions include a constant term. Standard errors robust to heteroskedasticity are in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. In Figure A-5, we report the results of the cross-country regressions by industry using the TiVA data for government and private purchases. The industry classification is more aggregated than in the ITPD-E database: there are 44 industries, 6 in the primary, 17 in the manufacturing and 21 in the service sector. We replicate the specification in Table A-2 for each industry and plot the coefficient on the protectionism dummy. The results are overall similar across government procurement and private purchases, confirming the findings on the aggregate data in Table A-2. The coefficients are more markedly negative when using the discrimination and preferential treatment indicators than the domestic sourcing one. Importantly, the coefficients for government procurement are never positive and significant, supporting the expectation that countries with protectionist laws in government procurement are less open to trade. Many of the industries with the strongest negative coefficients are in the service sector (e.g., utilities, education, postal services), although some manufacturing industries (e.g., motor vehicles, food products) are also among the ones where differences in import shares between protectionist and non-protectionist countries are more important. Therefore, the negative association between import shares and protectionism in government procurement law is rather present across different sectors. 32 Figure A-5. Results from cross-country regressions for government procurement and private purchases by industry 33 Notes: Each dot is the estimated coefficient on the protectionism indicator indicated in the title of the figure. Bars are 955 confidence intervals. The cross-country regression specification are as in Table A-2. The dependent variable is the import share in total government procurement (left-hand side column) and the import share in total private spending (right-hand side column). Government procurement and the private sector are defined in the text. The “protectionism” dummy equals the variables reported in Table A-5. Standard errors are robust to heteroskedasticity. As a last step in the econometric analysis, we estimate gravity regressions as described in section IV.B in the main text using the TiVA data and distinguishing between bilateral imports in government procurement and in private markets. As a check on the data similar to that of Table A-1, we report the results of the gravity regression using aggregate (the sum of government procurement and private purchases) bilateral imports and compare them with the ones obtained using the ITPD-E on the TiVA sample of countries (Table A-3). The coefficient on the interaction between the internal trade dummy and the composite protectionism indicator is positive and significant using both the ITPD-E and TiVA data. While the magnitudes are somewhat different, the results are qualitatively similar across the two data sources. The influence of protectionism in government procurement law on the border effect is stronger in the smaller TiVA sample than in the full sample used in the main analysis (Table 3). 34 Table A-3. Gravity regressions with ITPD-E and TiVA aggregate bilateral imports (1) (2) (3) (4) (5) (6) (7) (8) Dep. variable: bilateral TiVA data ITPD-E data imports Internal trade 2.632*** 3.365*** 2.246*** 1.540*** (0.160) (0.184) (0.198) (0.237) Log(distance) -0.578*** -0.436*** -0.578*** -0.426*** -0.509*** -0.411*** -0.512*** -0.404*** (0.040) (0.045) (0.040) (0.047) (0.046) (0.053) (0.047) (0.055) Common border 0.392*** 0.499*** 0.392*** 0.498*** 0.413*** 0.472*** 0.415*** 0.471*** (0.110) (0.107) (0.111) (0.106) (0.111) (0.107) (0.110) (0.106) Common language 0.381*** 0.190* 0.381*** 0.197** 0.471*** 0.288** 0.471*** 0.299*** (0.092) (0.101) (0.092) (0.100) (0.102) (0.112) (0.102) (0.110) Common legal origin 0.060 0.161*** 0.060 0.164*** 0.094 0.180*** 0.095 0.185*** (0.068) (0.061) (0.068) (0.061) (0.072) (0.066) (0.073) (0.067) Colony 0.319** 0.335* 0.318** 0.354** 0.221 0.236 0.215 0.245 (0.159) (0.174) (0.159) (0.173) (0.198) (0.211) (0.199) (0.209) WTO w/out GPA 0.667*** 0.303 0.666*** 0.307 0.283 0.782*** 0.266 0.785*** (0.187) (0.220) (0.190) (0.221) (0.238) (0.250) (0.237) (0.251) GPA 0.627** 0.030 0.627** 0.037 0.096 0.501* 0.066 0.507* (0.244) (0.243) (0.248) (0.241) (0.284) (0.264) (0.284) (0.263) PTA w/out procurement 0.168 0.750*** 0.165 0.783*** 0.586*** 0.975*** 0.594*** 1.005*** (0.117) (0.109) (0.118) (0.112) (0.127) (0.141) (0.128) (0.143) PTA with procurement 0.195** 0.188** 0.194** 0.201** 0.001 0.132 -0.014 0.138 (except EU) (0.081) (0.092) (0.081) (0.093) (0.089) (0.093) (0.090) (0.092) EU 0.735*** 0.920*** 0.734*** 0.943*** 0.849*** 0.987*** 0.843*** 1.003*** (0.175) (0.133) (0.176) (0.136) (0.179) (0.153) (0.178) (0.153) Internal trade x prot. 0.319*** 0.273** 0.320*** 0.252** 0.239* 0.383*** 0.247* 0.379*** Indicator (0.105) (0.111) (0.106) (0.115) (0.137) (0.141) (0.140) (0.143) Internal trade x WTO 0.504*** 0.390** 1.327*** 1.227*** w/out GPA (0.167) (0.170) (0.221) (0.217) Internal trade x GPA 0.018 -0.055 1.436*** 1.401*** (0.244) (0.242) (0.291) (0.287) Internal trade x EU 0.067 0.040 0.207 0.192 (0.128) (0.130) (0.143) (0.143) Internal trade x Ln(GDP -0.043 -0.126* -0.057 -0.159 cap) (0.057) (0.073) (0.081) (0.097) Internal trade x Ln(GDP) -0.234*** -0.240*** -0.219*** -0.226*** (0.042) (0.042) (0.053) (0.053) Internal trade x -0.007* -0.003 -0.012*** -0.008 Institutions WGI (0.004) (0.004) (0.004) (0.005) Obs 82764 82566 82764 82566 82755 82557 82755 82557 Notes: The dependent variable is the value of bilateral imports. Annual data from 2000 to 2018. The sample is given by the importing and exporting countries in the TiVA data. All columns include importer-year and exporter-year fixed effects. Columns (3), (4), (7) and (8) include interactions between the internal trade dummy and year dummies. Standard errors three-way clustered by undirected country-pair, importer and exporter in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. Table A-4 presents the estimate for public and private imports, and for each of the four protectionism indicators we use (three policy areas and the composite one). The evidence is consistent with the results from the cross-country regressions: the trade-reducing effect of protectionism in government procurement law is similar across public and private markets. If anything, the interaction coefficient is positive across all protectionism indicators for private purchases, whereas for government procurement it turns negative (suggesting that more protectionist countries have lower border effects) when we consider domestic sourcing and the combination of protectionism in all the three policy areas (lack of national treatment, preferential treatment and domestic sourcing). 35 Table A-4. Gravity regressions for government and private imports (1) (2) (3) (4) (5) (6) (7) (8) Government Private Pref. Dom. Overall Pref. Dom. Overall Discrimination Discrimination treatment sourcing indicator treatment sourcing indicator Internal trade 4.819*** 4.409*** 5.157*** 5.040*** 3.160*** 2.447*** 3.253*** 3.132*** (0.269) (0.269) (0.245) (0.226) (0.177) (0.209) (0.179) (0.182) Internal trade x prot. 0.256** 0.605*** -0.426*** -0.126 0.242** 0.912*** 0.123 0.279** Indicator (0.115) (0.167) (0.084) (0.115) (0.113) (0.155) (0.094) (0.109) Internal trade x WTO 0.965*** 0.821*** 0.878*** 0.807*** 0.482*** 0.348** 0.385** 0.514*** w/out GPA (0.223) (0.199) (0.197) (0.184) (0.161) (0.162) (0.164) (0.167) Internal trade x GPA -0.046 -0.103 0.030 -0.045 -0.085 -0.123 -0.050 0.040 (0.299) (0.278) (0.284) (0.268) (0.249) (0.245) (0.249) (0.243) Internal trade x EU 0.048 -0.059 -0.150 -0.132 0.125 0.025 0.008 0.074 (0.127) (0.113) (0.109) (0.117) (0.143) (0.124) (0.128) (0.126) Internal trade x Ln(GDP 0.192*** 0.154** 0.155** 0.128* -0.064 -0.090 -0.108* -0.068 cap) (0.067) (0.064) (0.061) (0.066) (0.060) (0.056) (0.056) (0.057) Internal trade x -0.466*** -0.452*** -0.433*** -0.439*** -0.215*** -0.213*** -0.206*** -0.217*** Ln(GDP) (0.049) (0.049) (0.047) (0.050) (0.042) (0.041) (0.043) (0.041) Internal trade x -0.010*** -0.009** -0.017*** -0.011*** -0.007* -0.005 -0.005 -0.007* Institutions WGI (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) Obs 82566 82566 82566 82566 82566 82566 82566 82566 Notes: The dependent variable is the value of bilateral imports for government procurement (columns (1) to (4)) and for private purchases (columns (5) to (8)). Annual data from the TiVA database from 2000 to 2018. All columns include importer-year and exporter-year fixed effects. Coefficients on the bilateral covariates are omitted for brevity (the specification is as in columns (1), (2), (5) and (6) of Table A-3). Standard errors three-way clustered by undirected country-pair, importer and exporter in parenthesis. * stands for statistical significance at the 10% level, ** at the 5% level and *** at the 1% percent level. Finally, Figure A-6 displays the coefficients on the interaction between the internal trade dummy and the protectionism indicators in our gravity regressions by industry in the TiVA data, distinguishing between government procurement and private imports. The results confirm the evidence from the aggregate data in the full sample. Industries in the services and primary sectors are on the two extremes – among the ones where protectionism in government procurement plays the strongest role, but also among the ones with counterintuitive negative coefficients. We also confirm that the results are the sharpest for preferential treatment, where the interaction coefficient is positive and significant for most industries. Importantly, the ranking of industries is similar across public and private purchases. 36 Figure A-6. Results from gravity regressions for government procurement and private purchases by industry 37 Notes: Each dot is the estimated coefficient on the interaction between the internal trade dummy and the protectionism indicator indicated in the title of the figure in gravity regressions. Bars are 955 confidence intervals. The gravity specification are as in Table A-4. The dependent variable is bilateral imports in government procurement (left-hand side column) and in total private spending (right-hand side column). Government procurement and the private sector are defined in the text. The “protectionism” dummies equal the variables reported in Table A-5. Standard errors are three-way clustered by undirected country-pair, importer and exporter. B. Additional tables Table A-5. Indicators for protectionism in the law on government procurement Country Discrimination Dom. sourcing Pref. treatment Protectionism AFG 1 0 1 0 AGO 1 1 1 1 ALB 1 0 1 0 ARG 0 1 1 0 ARM 0 0 0 0 ATG 1 0 0 0 AUS 0 0 1 0 AUT 0 0 1 0 AZE 1 1 1 1 BDI 0 1 1 0 BEL 0 1 1 0 BEN 0 0 1 0 BGD 0 1 1 0 BGR 1 0 1 0 BHS 1 1 1 1 BIH 1 0 1 0 BLZ 1 0 1 0 BOL 0 1 1 0 BRA 1 1 1 1 BRB 1 0 0 0 BRN 0 0 0 0 BTN 0 1 1 0 BWA 0 1 1 0 CAN 0 0 1 0 CHE 0 1 1 0 CHL 0 0 1 0 CHN 0 1 1 0 CIV 0 0 1 0 CMR 0 1 1 0 COD 1 0 1 0 COG 0 0 1 0 COL 0 1 1 0 CPV 1 1 1 1 CRI 0 1 1 0 CUB 0 0 1 0 CYP 0 0 1 0 CZE 0 1 1 0 38 DEU 0 0 1 0 DMA 1 1 1 1 DNK 0 0 1 0 DOM 0 0 0 0 DZA 0 1 1 0 ECU 0 1 1 0 ESP 0 0 1 0 EST 1 0 1 0 ETH 0 1 1 0 FIN 0 0 1 0 FJI 1 0 1 0 FRA 0 1 1 0 GAB 0 1 1 0 GBR 0 0 1 0 GEO 1 0 0 0 GHA 1 1 1 1 GRC 0 1 1 0 GRD 1 1 1 1 GTM 0 0 0 0 GUY 1 1 1 1 HKG 0 0 0 0 HND 0 1 1 0 HRV 0 1 1 0 HTI 0 1 1 0 HUN 0 0 1 0 IDN 1 1 1 1 IND 1 1 1 1 IRL 0 0 1 0 IRN 1 1 1 1 IRQ 0 0 1 0 ISL 0 0 1 0 ISR 0 1 1 0 ITA 0 0 1 0 JAM 0 1 1 0 JPN 1 0 1 0 KAZ 1 1 1 1 KEN 0 1 1 0 KHM 0 1 1 0 KOR 1 0 1 0 LAO 0 1 1 0 LBR 1 1 1 1 LKA 1 1 1 1 LTU 0 0 1 0 LUX 0 1 1 0 LVA 1 0 1 0 MAR 0 0 1 0 MDA 1 0 0 0 MDG 0 0 1 0 MEX 0 1 1 0 MLT 0 0 1 0 MMR 1 1 1 1 MNE 0 0 1 0 MNG 1 1 1 1 MOZ 0 1 1 0 MRT 0 0 1 0 MUS 1 1 1 1 MWI 1 1 1 1 MYS 0 1 1 0 NAM 1 1 1 1 NGA 1 1 1 1 NIC 0 0 . 0 NLD 0 0 0 0 NOR 0 0 1 0 NZL 0 0 0 0 PAK 0 1 1 0 PAN 0 0 1 0 PER 0 0 0 0 39 PHL 1 1 1 1 PNG 0 1 1 0 POL 1 1 1 1 PRT 0 0 1 0 PRY 0 1 1 0 ROU 0 0 1 0 RUS 1 1 1 1 RWA 1 1 1 1 SAU 0 1 1 0 SEN 0 1 1 0 SGP 0 0 0 0 SLE 1 1 1 1 SLV 0 0 0 0 SSD 1 1 1 1 SVK 0 0 1 0 SVN 0 1 1 0 SWE 0 0 1 0 SWZ 0 1 1 0 SYC 1 0 1 0 TGO 0 1 1 0 THA 1 1 1 1 TKM 1 1 0 0 TLS 0 1 1 0 TTO 1 1 0 0 TUN 0 1 1 0 TUR 1 1 1 1 TWN 1 1 1 1 TZA 1 1 1 1 UGA 0 1 1 0 UKR 0 0 0 0 URY 0 1 1 0 USA 1 1 1 1 UZB 0 1 1 0 VEN 0 1 1 0 VNM 1 1 1 1 ZAF 1 1 1 1 ZMB 0 1 1 0 ZWE 0 1 1 0 Average 0.366 0.585 0.879 0.239 Notes: The indicators in 2nd to 4th columns are defined as described in the main text (section III.A). The “protectionist” indicator equals one if the three indicators in columns 2 to 4 equal one, and zero otherwise. 40 Table A-6. Results from cross-country regressions by industry Industry Discrimination Pref. treatment Dom. sourcing Protectionism Distilling rectifying & -0.197 -0.272 blending of spirits (0.100) (0.137) Furniture -0.186 (0.091) Mining of lignite -0.183 (0.089) Insulated wire and -0.161 cable (0.078) Parts/accessories for -0.150 automobiles (0.066) Spices -0.477 (0.141) Tobacco products -0.521 -0.359 -0.630 (0.189) (0.164) (0.291) Automobile bodies -0.323 (0.084) Plastic products -0.322 -0.282 (0.115) (0.121) Wooden containers -0.306 (0.138) Fresh fruit -0.561 (0.280) Rice (raw) -0.523 (0.246) Pharmaceuticals -0.109 (0.050) Electronic valves tubes -0.066 (0.029) Other mining and -0.439 quarring (0.173) Processing/preserving -0.269 of fruit & vegetables (0.123) #(share) significant at 7(4.5%) 43(27.6%) 6(3.9%) 25(16%) 5% Average (if -0.147 -0.174 -0.277 -0.192 significant): - Primary -0.183 -0.341 -0.542 -0.127 - Manuf. -0.141 -0.166 -0.144 -0.203 - Services . . . . Notes: The top panel reports for each column the largest five coefficients estimated from industry-specific cross-country regression with controls (specifications in even-numbered columns in Table 1). The first column reports the name of the industry, following the classification in the ITPD-E database. We select only coefficients that are significantly different from zero at 5% level. Standard errors robust to heteroskedasticity are in parenthesis. The middle panel reports the share of industry-specific coefficients that are different from zero at the 5% level. The bottom panel reports the average across the industry-specific coefficients conditional on those being statistically significant, overall and by sector. 41 Table A-7. Results of gravity regressions by industry Industry Discrimination Pref. treatment Dom. sourcing Protectionism Processing of nuclear 7.921 4.539 fuel (1.216) (2.130) Office accounting and 4.151 computing machinery (0.975) Publishing of recorded 3.223 3.397 media (0.552) (0.522) TV and radio receivers 3.137 (0.789) Optical instruments & 2.810 photographic equipment (0.781) Eggs 4.278 (0.589) Mining of iron ores 3.750 3.482 -0.630 (0.838) (0.773) (0.291) Other meats -3.279 -2.894 (0.979) (0.892) Aircraft and spacecraft 3.084 -0.282 (0.939) (0.121) Machine tools 2.977 (0.786) Cigarettes -4.893 -3.202 (0.933) (0.531) Rice (raw) -4.026 (0.859) Beverages nec -3.601 (0.570) Other personal services -3.108 (1.059) Weapons -2.857 (0.653) #(share) significant at 50(33%) 61(40%) 30(20%) 59(38.6%) 5% Average (if 0.207 1.280 -0.440 -0.457 significant): - Primary -0.644 1.645 -1.212 -0.769 - Manuf. 0.640 1.225 -0.096 -0.141 - Services -1.637 1.126 . -1.912 Notes: The top panel reports for each column the largest five coefficients estimated from industry-specific gravity regression with bilateral and other importer-specific control variables (specifications in odd-numbered columns in Table 4). The first column reports the name of the industry, following the classification in the ITPD-E database. We select only coefficients that are significantly different from zero at 5% level. Standard errors are three-way clustered by undirected country-pair, importer and exporter. The middle panel reports the share of industry- specific coefficients that are different from zero at the 5% level. The bottom panel reports the average across the industry-specific coefficients conditional on those being statistically significant, overall and by sector. 42