Policy Research Working Paper 10339 Natural Resource Dependence and Monopolized Imports Rabah Arezki Ana Fernandes Federico Merchán Ha Nguyen Tristan Reed Development Economics Development Research Group March 2023 Policy Research Working Paper 10339 Abstract Countries with greater commodity export intensity have measures, and tariff evasion are mechanisms that concen- more concentrated markets for imported goods. Import trate import markets. These results suggest a novel channel market concentration is associated with higher domestic for the resource curse stemming from the monopolization prices, suggesting that markups due to greater concentra- of imports. tion outweigh any potential cost efficiency. Tariffs, non-tariff This paper is a product of the Development Research Group, Development Economics. 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 afernandes@worldbank.org and treed@worldbank.org. 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 Natural Resource Dependence and Monopolized Imports ´n, Ha Nguyen, Tristan Reed∗ Rabah Arezki, Ana Fernandes, Federico Mercha Originally published in the Policy Research Working Paper Series on March 2023. This version is updated on June 2024. To obtain the originally published version, please email prwp@worldbank.org. JEL Classification: O12; O13; F14; F10; L12. Keywords: imports, market concentration, natural resources, resource curse. ∗ Rabah Arezki is Director of Research at the French National Research Center (CNRS) and a Senior Fellow at the Harvard Kennedy School, Ana Fernandes and Tristan Reed are Lead Economist and Economist at the World Bank Development Research Group, Ha Nguyen is Economist at the International Monetary Fund (IMF), and Federico Merch´ an is a Junior Researcher at the Institute for the World Economy (IFW). We thank Leila Baghdadi, Olivier Blanchard, Shanta Devarajan, Simeon Djankov, Jeffrey Frankel, Caroline Freund, Penny Goldberg, Jean Imbs, Daniel Lederman, Rick van der Ploeg, Bob Rjikers, Gregoire Rota- Graciozi, and Tony Venables for valuable comments and suggestions, as well as participants at the 2nd Annual Central Bank Conference on Development Economics in the Middle East and North Africa. Mayra Monroy, Jan Oledan, Gaston Nievas, and Paula Suarez provided excellent research assistance. This paper benefited from support from the Umbrella Facility for Trade trust fund financed by the governments of the Netherlands, Norway, Sweden, Switzerland and the United Kingdom and the Multi-Donor Trust Fund (MDTF) from the World Bank’s Middle East and North Africa Region. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the World Bank and the IMF, the Executive Directors of the World Bank and the IMF or the governments they represent. 1 Introduction Dependence on natural resources for exports creates a variety of macroeconomic challenges known collectively as ‘the resource curse’ (Sachs and Warner, 2001; Van der Ploeg, 2011; Frankel, 2012; Venables, 2016; Arezki, Ramey and Sheng, 2017). One challenge is the so- called Dutch disease wherein a natural resource discovery or price appreciation is accompa- nied by a real exchange rate appreciation, which in turn shrinks the non-resource tradable sector leading to de-industrialization. Such appreciation also leads to an increase in the size of the import market.1 Another challenge faced by resource-dependent economies is rent- seeking wherein natural resource rents controlled by the state increase the return to state capture, leading to inefficient policy choices in the absence of strong institutions. In principle, both the Dutch disease and rent-seeking challenges could interact to form an under-explored “import” channel for the resource curse. Foreign exchange receipts from natural resources combined with real exchange rate appreciation expand the domestic demand for imports, increasing the domestic import market. By making the import market larger, natural re- sources raise the return to effort by importers towards capturing the state and using trade or industrial policy to shield them from competition. Yet, existing theoretical models of state capture in natural resource-dependent economies (Tornell and Lane, 1999; Robinson, Torvik and Verdier, 2014) do not emphasize profits in the import market as a source of rents. Anecdotal evidence is consistent with this import monopolization effect, as the wealth of many billionaire businesspeople in natural resource-dependent economies is linked to profits in import markets. Prominent billionaires Femi Otedola in Nigeria, Abdul Latif Jameel in Saudi Arabia, and Igor Kesaev in the Russian Federation accumulated their wealth respec- tively as a fuel importer, an exclusive distributor for a car manufacturer, and an importer of alcohol, cigarettes, and food.2 This paper moves beyond anecdotes and provides systematic 1 Harding and Venables (2016) document a positive association between resource export revenues and the size of the import market. In our sample, a bivariate regression confirms that a 1 percentage point increase in the share of commodities in total merchandise exports is associated with a 0.6 percent increase in the value of the import market. 2 See Freund (2016) for an account of the origin of billionaires’ wealth in emerging markets. 2 evidence that natural resource dependence is associated with monopolization of imports, and that trade policy is a mechanism behind this association. The term ‘monopolization’ is used to describe a shift in market structure toward one that is more concentrated. The analysis exploits a novel database of all firm-level import transactions in 53 devel- oping economies. These data reveal that natural resource-dependent economies have more concentrated markets for imported products. This main result is illustrated in Figure 1a, which shows a positive association between countries’ commodity exports as a share of total merchandise exports and the weighted average Herfindahl–Hirschman index (HHI) across their imported product markets. The HHI for an imported product market is the sum of the squared market shares of every firm importing that product. Econometric estimates show this relationship is robust to a variety of controls including country-product and product- year fixed effects. An alternative specification shows exogenous increases in world commodity prices also increase import market concentration, suggesting the relationship is causal. Traditional models of Dutch disease emphasize the increase in the price of non-tradable goods relative to the price of tradable goods hence the appreciation of the real effective exchange rate. Yet, these models ignore elements of market structure and the presence of markups. These elements can affect prices over and above the effect on relative prices stem- ming from the traditional Dutch disease. Our results suggest that prices are elevated due to higher markups ensuing from monopolistic or oligopolistic pricing by importers. Data on import unit values and domestic price levels show that import market concentration can account for higher prices. This result is not obvious ex-ante: if higher market concentra- tion is associated with a higher fixed cost but lower marginal cost of importing, import market concentration could be associated with lower prices, even if it contributes to higher markups. Along these lines, contributions by Edmond, Midrigan and Xu (2023) and Aghion, Bloom, Blundell, Griffith and Howitt (2005) show that increased markups stemming from high market concentration may (though need not) harm welfare. Higher prices associated with concentration could shape firms’ international input sourcing decisions, as recently ana- 3 as, Fort and Tintelnot, 2017; Goldberg and Reed, 2023). lyzed using United States data (Antr` To the extent that import monopolization raises costs of input procurement in global value chains, it may impede countries’ efforts to diversify exports away from natural resources. We provide evidence suggesting trade policy is a channel for import monopolization in natural resource-dependent economies. Figure 1b shows the commodity export share is positively associated with higher import tariffs. We also find the commodity export share is associated with greater use of non-tariff measures (NTMs) such as quotas, that limit entry into imports. Trade protection increases import market concentration through direct and indirect mech- anisms. The direct mechanism is that trade protection, ceteris paribus, both reduces demand and raises the cost of importing, which could lead firms to exit and increase concentration. Our evidence is consistent with this mechanism: imported product markets with higher tar- iffs are more concentrated, conditional on country-year and product-year fixed effects. The country-year fixed effects rule out this evidence being driven by differences in national mar- ket size or institutions, and the product-year fixed effects rule out this evidence being driven by differences in the global product market, like technology or the product’s share in the import basket of a typical developing country.3 The indirect mechanism is tariff avoidance. Certain importers may be able to avoid paying tariffs, either by securing legal exemptions or by illegal tariff evasion. They thus achieve an asymmetric cost advantage that allows them to sell imports more cheaply than firms that cannot avoid tariffs, expanding their market share and increasing market concen- tration. The seminal literature on rent-seeking suggested as much, arguing tariff avoidance could explain the persistence of high tariffs, as elites that enjoy the advantage of avoidance are a constituency in favor of tariffs remaining high (Tullock, 1967; Krueger, 1974). 3 Our data, which begin in 1997, succeed major tariff reforms in most countries, so we are unable to exploit within country-product variation in trade policy that would allow inclusion of country-product fixed effects in this analysis. 4 Our evidence is also supportive of this indirect mechanism, but only in countries with certain types of commodity exports. Isham, Woolcock, Pritchett and Busby (2005) argue rent-seeking associated with the resource course is greater in fuel exporting economies be- cause fuel extraction is point-based, meaning revenues typically transit directly through government coffers, in contrast with economies exporting commodities with a more diffuse production base. Consistent with this, we find that control of corruption measured by expert surveys is weaker in economies dependent on fuel exports compared to economies specializing in exports of ores and metals or food commodities. In our sample, the association between tariffs and import market concentration is stronger for economies that are dependent on fuel exports than for those that are dependent on ores and metals or food. Using a measure of tariff evasion based on the underinvoicing of imports relative to exports reported by part- ner countries, we find a positive relationship between the tariff rate and tariff evasion in economies dependent on fuel exports, but not in economies dependent on other commodity exports. This evidence suggests that tariff evasion explains import market concentration especially in fuel exporting economies. Our paper is the first to systematically explore differences across countries in import mar- ket structure, contributing to several literatures beyond that about the resource curse. While the export sector has been the traditional focus of the trade and development literature, in developing countries the value of imports is about as large as that of exports, and many exported goods are produced using imported inputs (UNCTAD, 2021). We identify patterns in import market structure that contrast starkly with those in studies examining export market structure. Fernandes, Freund and Pierola (2016) use the same customs transactions data on the export side to document that higher-income economies have more exporting firms, but also more concentrated export markets dominated by “superstars,” or firms with especially large market shares, whose characteristics are described by Freund and Pierola (2015, 2016).4 The pattern in import markets is the opposite. Higher-income economies 4 Freund and Pierola (2015) show national revealed comparative advantage is shaped by the presence of 5 have less concentrated import markets, independent of their commodity export intensity. Our paper provides general insight into differences in market concentration across coun- tries. Leone, Macchiavello and Reed (2024) describe how concentration leading to high markups has raised prices in Africa’s domestic cement industry, and conclude the reason for concentration is small national market size, rather than entry costs that are higher in Africa. In contrast, the present paper provides evidence of entry costs in importing that are unique to commodity export intensive economies and can account for higher costs in these economies. The remainder of the paper is organized as follows. Section 2 describes the data and measurement of import market concentration. Section 3 shows the main result in Figure 1a is robust to a battery of controls and alternative regression specifications. Section 4 presents evidence that tariffs, non-tariff measures, and tariff evasion are mechanisms that concentrate import markets. Section 5 documents the positive association between domestic prices and import market concentration. Section 6 concludes. 2 Data and measurement Import Market Concentration. Import market concentration is measured using a novel database of all firm-level import transactions recorded by customs authorities in 53 coun- tries, geographically and institutionally diverse, and broadly representative of middle-income economies (see Table S1). The database has the same source as the World Bank Exporter Dynamics Database described by Fernandes et al. (2016) but includes import rather than export transactions. The sample period covers 1997-2021 but with different year coverage for each country. We eliminate observations in HS27 (oil, petroleum, natural gas, and coal) as their trade is not uniformly reported across countries’ customs data. Country-year total superstar exporters. Gaubert, Itskhoki and Vogler (2021) discuss the policy implications of such “granular” comparative advantage in exports. Our evidence highlights that import markets can also be granular, with implications for the price level. 6 non-oil imports in our data are very similar to the corresponding total non-oil imports re- ported by COMTRADE (the average difference across country-years is 5.6%). COMTRADE data is accessed through WITS (World Bank, 2024b). Measuring import market concentration requires defining a relevant market, or the set of products over which the firms in question have market power. While relevant market definition is the object of intense debate in antitrust litigation, a general principle is that it should include the set of goods that are close substitutes for the same set of consumers es, 2009). Benkard, Yurukoglu and Zhang (2021) note that economic census (Davis and Garc´ data, whereby firms are classified into industries, are collected at the point of production rather than consumption, and so may be less useful for analyzing the relation between market concentration and market power. In contrast, the Harmonized System (HS) product categories defined by the United Nations used in trade data classify goods with a similar end-use, and so are conceptually like the relevant product markets in antitrust analysis. Following Fernandes et al. (2016), we use a time-consistent consolidated classification that concords and harmonizes product codes across the HS revisions present in the raw data.5 Our baseline measure of import market concentration is the Herfindahl-Hirschman Index (HHI) or J 2 Mc,j,i,t HHIc,i,t = 100 × j Mc,i,t where Mc,j,i,t is the import value of firm j in country c of product i in year t, J is the number of firms and Mc,i,t the total import value in the relevant import country, product market, and year. Values are measured including cost of freight and insurance (CIF) in United States dollars. Table S1 reports average and median HHI across HS 6-digit products by import country. We consider alternative measures of concentration: the share of the largest importing firm or the largest four importing firms, the number of importing firms, and the number of source countries, all defined by import country, product market and year. We also 5 The consolidation incorporates HS revisions 1996 through 2017. A description of the methodology is available upon request. 7 use unit values defined as the ratio of import value to import weight by country-product- year. Moreover, we measure the import share for a specific set of firms, those that are also exporters of commodities, identified based on a database of all firm-level export transactions recorded by customs authorities in the same 53 countries. Natural resource dependence. We employ two measures capturing natural resource dependence: (i) the share of commodities in total merchandise exports from the World Development Indicators (World Bank, 2024a); and (ii) an index relying only on fluctuations in world commodity prices (Gruss and Kebhaj, 2019).6 Means and standard deviations of these variables are reported in Table S2. Commodities and natural resources are defined broadly to include food (e.g., animal and vegetable oils and fats, beverages, fruits and vegetables, live animals); fuels (e.g., mineral fuels, lubricants, related materials); and ores and metals (e.g., crude fertilizers, metalliferous ores, non-ferrous metals). Measure (i) captures both price and quantity variation in commodity exports, and is our preferred measure of natural resource dependence. In some specifications we separate this into exports of different types of commodities: food, fuels, and ores and metals. Measure (ii) captures only price variation. The advantage of this measure is that it is plausibly exogenous, since it excludes quantities determined by local production costs and world prices should not be affected by local exports as economies in our sample are small relative to the commodity markets they export to. A disadvantage of this measure is that price variation alone does not capture the dependence of an economy on natural resources for export revenue. In sum, there is a trade-off between ”quantity” (export share) and ”quality” (price index) of variation in natural resource dependence. 6 K For country c in year t this index is given by k=1 log(Pk,t )ωc,k where Pk,t is the world price of commodity k in year t, and ωc,k is the weight given by the average value of commodity k ’s exports as a share of GDP across the 1980-2020 period. The index is scaled for each country so 100 equals the price index in 2012. 8 Prices. Prices are measured in two ways: (i) import unit values (import value divided by import weight); and (ii) national average prices of products collected by the International Comparison Program (ICP) (World Bank, 2011, 2017). Since import weight is not available for Chile, India, Mexico, and Vietnam, the results for unit values are based on a smaller sample. Other variables. As measures of trade policy, applied tariff rates and NTMs for HS 6-digit products are sourced from TRAINS (UNCTAD, 2024). For tariff rates we use the simple average across origin countries of the applied tariff within an importing country-product-year. For NTMs, we focus on Chapters E “Non-automatic import licensing, quotas, prohibitions, quantity-control measures” and F “Price-control measures, including additional taxes and charges” as defined by UNCTAD (2019). We define an indicator that equals one if any of those NTMs are present within an importing country-product-year and zero otherwise. As controls, in some specifications we include log real GDP per capita at purchasing power parity and the most recent Gini coefficient from the World Development Indicators (World Bank, 2024a), and indexes of control of corruption and regulatory quality (Kaufmann, Kraay and Mastruzzi, 2010). Means and standard deviations of these variables are reported in Table S2. 3 Main result Figure 1a showed a significant correlation between the commodity export share and im- port market concentration. We demonstrate that this relationship is robust to a battery of controls, using the following regression HHIc,i,t = αc,i + δi,t + β1 ExpComc,t + βX Xc,t + ϵc,i,t (1) where αc,i is a country-product market fixed effect that captures unobserved market charac- 9 teristics that may explain concentration (e.g., market size, consumer preferences) and δi,t is a product-year fixed effect that controls for global supply and demand factors in the product market that do not vary across countries (e.g., technological fixed costs, per-unit good value, logistics network requirements). The independent variable of interest is ExpComc,t defined as either the commodity share of exports or the export commodity price index, measured as differences from the sample median. Xc,t is a vector of country-year controls used in some specifications, also measured as differences from the sample median, and ϵc,i,t is an error. This specification has several differences compared to Figure 1a. First, this specification focuses on the association of commodity exports with the level of HHI rather than the log. This approach is selected so results can be compared to a benchmark of how large an increase in HHI is presumed to increase market power. United States Federal Trade Commission (FTC) and Department of Justice (DOJ) guidelines presume that an increase in the HHI of 100 will increase market power in a highly concentrated market.7 A market with HHI in excess of 1,800 is considered highly concentrated. Second, compared to Figure 1a, the regression weighs product markets equally rather than by value, so as not to focus primarily on highest value products (e.g., autos) and instead characterize the average product market. Results for all products. Table 1, Panel A reports estimates of Equation 1 on a sample pooling across all imported products. Column 1 reports the raw correlation of the commodity export share (less its median) and HHI, with no fixed effects or controls. The mean dependent variable is 4,219, indicating on average highly concentrated markets according to FTC and DOJ guidelines. Here, β1 = 901 (standard error = 132). At the median, a one standard deviation increase in the commodity export share of 0.24 is associated with an increase in HHI equal to 901 × 0.24 = 216, more than sufficient to increase market power according 7 See United States Federal Trade Commission and Department of Justice (2023). Previous FTC and DOJ guidelines presumed an increase in HHI of 200 would increase market power. Nocke and Whinston (2022) conclude that an increase in HHI between 100 and 200 can reduce consumer welfare given small efficiency gains from a horizontal merger that increases concentration. 10 to FTC and DOJ guidelines. Column 2 reports results adding product-year fixed effects, where β1 = 1, 238 (169). The increase in the coefficient suggests that commodity exporting economies are likely to import products whose markets are less concentrated due to global factors, such as technology with lower fixed costs. Column 3 adds country-product fixed effects, which isolates only variation in HHI due to within-country over-time variation in the commodity export share. Here β1 = 438 (228). The differences between Columns 1, 2, and 3 suggest that some, but not all, of the association between HHI and natural resource dependence is due to differences in product mix. Column 4 adds several country-year control variables: (i) the log real GDP per capita, a proxy for market size and level of development; (ii) the Gini coefficient, a measure of inequality, which could also shape market size; and experts’ perceptions of (iii) control of corruption; and (iv) regulatory quality. The estimate of β1 is robust to their inclusion: it is 358 (170), which is lower than, but not statistically distinguishable from, the coefficient in Column 3 at standard significance levels. The coefficients on the controls are informative. Higher GDP per capita is associated with lower import market concentration. This is consis- tent with richer countries having larger import markets and therefore being able to sustain more entrants. This result is in contrast to the findings of Fernandes et al. (2016) that exporter concentration within a country rises with GDP per capita and suggests potential scale economies in importing. Higher control of corruption and regulatory quality are also associated with lower import market concentration, suggesting that weak institutions can concentrate markets. Most surprising is the result that greater inequality is associated with lower import market concentration. This may reflect the hypothesis that in middle-income countries imports are consumed by wealthier individuals, whereas domestic production is consumed by the poor and the middle class. Hence, inequality could increase the market size for imports, though we do not test this hypothesis in this paper. In subsequent analysis, we do not include these controls in our main specification as they could be “bad controls” in the Angrist and Pishcke sense: outcomes of the independent 11 variable that cause the dependent variable. For instance, weak control of corruption or low quality regulation, which enable tariff avoidance, could be mechanisms through which commodity export dependence leads to import concentration. To the extent that income and inequality proxy for institutional weakness, they are also subject to this critique. From this perspective, it is difficult to interpret the relationship between natural resource dependence and import concentration conditional on these country-year factors. Columns 5 and 6 use the export commodity price index in the place of the commodity export share. In Column 5, which includes country-product and product-year fixed effects but no controls, β1 = 1, 050 (614). This implies that a one standard deviation increase in the price index of 0.03 is associated with an increase in HHI of 1, 050 × 0.03 = 31.5, much smaller than a one standard deviation increase in the commodity export share, and not economically significant. In Column 6, an economically similar result obtains when adding the country- year controls. These results suggest that economically relevant variation in the HHI is driven by the quantity of commodity exports rather than their price alone. Nonetheless, these results exploiting exogenous international price variation give us confidence the association between natural resource dependence and import market concentration is causal. Results by product category. Table 1, Panel B explores the heterogeneous association between commodity export intensity and import market concentration by splitting the sam- ple between goods with different end uses: capital goods, consumption goods, and primary goods according to the United Nations Broad Economic Categories. Examples of primary goods are iron ore, raw sugar, soybeans, and wheat; examples of capital goods are transmis- sion apparatuses, data processing machines, and airplanes; examples of consumption goods are medicaments, small vehicles, and televisions. Columns 1, 2, and 3 report the associa- tions of the commodity export share and HHI of imported primary, capital, and consumption goods markets respectively; Columns 4, 5, and 6 report the associations of the export com- modity price index and HHI of those same goods markets. For the export commodity share, 12 there is no significant heterogeneity, but for the export commodity price index there is signif- icant heterogeneity, with the association being much stronger for primary and consumption goods. Results for alternative measures of import concentration. To complement our base- line results using HHI, we present in Table 2 estimates of Equation 1 using as dependent variables alternative import market concentration measures. The results show that commod- ity exports result in higher concentration of imports in the top firm (column 1) or the top four firms (column 2). Focusing on different extensive margins of imports, the evidence im- plies that countries more dependent on commodity exports have significantly lower numbers of importing firms (column 3) and import sources (column 4). As a manifestation of import concentration, we consider unit values per market and find them to be higher for countries with higher commodity export shares or price indexes (column 5). Finally, higher commod- ity export quantities or prices may increase the import market share of firms that are also commodity exporters, and thus import concentration, though the effect is not statistically significant (column 6). 4 Trade policy mechanisms for import monopolization Figure 1b showed that commodity export dependent economies have greater trade protection. There are multiple potential reasons for this, including industrial policy intended to develop infant industries and vested interests that gain commercial advantages from avoiding tariffs. Whatever the reason for this pattern, in this section we present evidence that trade protection can concentrate import markets. We consider two protectionist trade policies: tariffs and non-tariff measures imposed on imports. Tariffs. Tariffs could cause concentration through three mechanisms: 1. Demand decrease: When passed through to consumers, tariffs raise prices and decrease 13 demand, potentially leading to exit of importing firms and increased concentration of market shares.8 2. Symmetric cost increase: Tariffs raise working capital requirements and financing costs, since they must be paid upon import before goods are sold, potentially leading to exit of importing firms and increased concentration of market shares. 3. Asymmetric cost increase due to tariff evasion: In a weak governance environment and in the presence of corruption, tariff evasion allows evading importing firms to increase their costs by less than others, increasing their market share relative to firms that do not evade. We test for these mechanisms using a mediation analysis that evaluates the association of tariffs and import market concentration, and how that is mediated by commodity exports. Table 3 reports regressions of the following form: yi,c,t = δc,t + δi,t + β2 Ti,c,t + β3 (Ti,c,t × ExpComc,t ) + υi,c,t (2) where yi,c,t is an outcome, either HHI or a measure of tariff evasion, the latter to identify the role of corruption; Ti,c,t is a measure of trade policy, either the tariff rate ∈ [0, 1] or a dummy variable ∈ {0, 1} indicating the presence of an NTM; δc,t is a country-year fixed effect subsuming national characteristics, including market size and commodity export dependence of country c; δi,t is a product-year fixed effect capturing global supply and demand for product i; and υi,c,t is an error. Country-product fixed effects are excluded so the identifying variation is across products within countries. The coefficients of interest are β2 , the direct association of trade policy and import market concentration, and β3 , the mediated association. The estimate in Column 1 of Table 3 is consistent with the demand and symmetric cost increase channels, as β2 = 3,514 (1,019), indicating higher tariffs are associated with 8 We confirm in our sample that import markets shrink as they become more concentrated: the elasticity of import value to HHI, conditional on country-product fixed effects, is -0.24 (0.002). 14 more concentrated markets. This relationship is economically significant, with a standard deviation increase in the tariff rate of 0.05 (or 5 percentage points) raising HHI by 3, 514 × 0.05 = 176, above the amount presumed to increase market power in current FTC and DOJ guidelines. We test for the asymmetric cost increase channel by measuring tariff evasion directly and exploiting variation in the type of commodity export that is associated with rent-seeking. Isham et al. (2005) find that rent-seeking associated with the resource curse is greatest when countries export ‘point-based’ resources, whose revenues transit directly through government coffers, as opposed to ‘diffuse’ resources whose revenues flow to many small holders. Oil is the quintessential point-based commodity whose extraction is often controlled by the state and is associated with state capture (see, e.g., Ross, 2012). The anecdotes of import monopolies cited in the introduction are in oil-exporting states. In contrast, production of food crops is typically diffuse. Ores and metals are an ambiguous case: industrial extraction can be capital-intensive and thus point-based, but labor-intensive artisanal mining with diffuse ownership can account for a substantial portion of output in some countries due to variation in geography (Rigterink, Ghani, Lozano and Shapiro, 2022). Consistent with these arguments, Table S3 shows there is less control of corruption as the commodity export share increases, but this relationship is strongest for countries with larger fuel exports, weakest for countries with larger food exports, and of intermediate magnitude for countries with larger ore and metals exports. Table S3 also shows there is no statistically significant relationship between changes in the commodity export price index and control of corruption, suggesting that commodity export quantities generate important variation in corruption beyond price. This finding motivates our focus on the commodity export share as an independent variable in this analysis. Column 2 of Table 3 reports estimates of Equation (2) where ExpComc,t is separated into three variables, each indicating the share of exports in these three commodity types, measured as differences from the median. In this specification β2 = 1, 794 (1, 073), which im- 15 plies that a standard deviation increase in the tariff rate increases HHI by 1, 794 × 0.05 = 90 points. The coefficient on the interaction between tariff and fuel export share is positive f uel β3 = 8, 696 (1,907) indicating that the positive association between tariffs and concen- tration is much stronger in fuel exporting economies: at the median, a 5 percentage point increase in the tariff rate (the standard deviation) and an 18 percentage point increase in fuel export share (the standard deviation) will raise HHI by about 8, 696 × 0.05 × 0.18 = 78 points. The interaction term for ores and metals exports is positive, though much smaller, and the interaction term for food exports is negative. The result is clear that tariffs contribute to import market concentration, but especially in fuel exporting economies. To confirm that the interaction effect is due to greater tariff evasion in fuel exporting economies, we use a proxy for tariff evasion as the dependent variable in the same regression. We use the measure of Fisman and Wei (2004, note 6) based on import underinvoicing: (exportsw,i,c,t − importsc,i,t ) evasion gapc,i,t = (exportsw,i,c,t + importsc,i,t ) where exports are total exports by all countries w of product i to destination country c (one of our importing countries) in year t reported in COMTRADE, and imports are as reported in our customs data. Results are reported in Column 3 of Table 3. The interaction coefficients follow a similar pattern as in Column 2, where the coefficient on the interaction f uel between tariff and fuel export share is positive β3 = 0.36 (0.17), indicating greater evasion, but the coefficients on the interaction between tariff and ores and metals export share and food export share is negative.9 Higher tariffs are associated with higher values of the proxy for tariff evasion in fuel exporting economies, but not in ores and metals or food exporting 9 The import underinvoicing proxy for tariff evasion has been found to be positively correlated with higher tariffs (Bhagwati, 1964; Fisman and Wei, 2004; Mishra, Subramanian and Topalova, 2008; Javorcik f uel and Narciso, 2017). In our sample, the value of β2 + β3 = −0.11 (0.18) so a positive association between tariffs and evasion cannot be rejected in fuel exporting economies (i.e., the upper bound of the 95% confidence interval of this term is 0.24). 16 economies. This result suggests that while the demand and symmetric cost increase mecha- nisms of tariffs explain import market concentration in all commodity exporting economies, the asymmetric cost increase mechanism through tariff evasion has explanatory power in fuel exporting economies. This result is consistent with less control of corruption in fuel exporting economies. Non-tariff measures. NTMs, in particular measures that restrict entry or pricing among importers, namely licenses, quotas, prohibitions, and price-control measures, including addi- tional taxes or fees paid at customs, are other forms of trade protection that could influence import market concentration. The mean of our NTM indicator is 0.42, indicating these NTMs are quite common. Regressing this indicator on the commodity export share (less its median) in a linear probability model yields a positive coefficient on the commodity export share equal to 0.29 (0.12), indicating NTMs are more common in countries with a higher commodity export share. In our sample, at the median, an increase in the commodity export share by one standard deviation, or 24 percentage points, increases the likelihood of an NTM by 0.29 × 0.24 = 0.07 or 7 percentage points. Columns 4, 5 and 6 of Table 3 replicate Columns 1, 2 and 3, but where the measure of trade protection is the NTM indicator. In Column 4, as in Column 2, the coefficient β2 is positive, though quantitatively the effect is smaller. The coefficient β3 is also positive. In Column 5, we again separate out the different types of commodity exports. The interactions of the NTM indicator with fuel export share and the ores and metals export share are positive, while the interaction with food exports is negative. These results indicate that NTMs can explain import market concentration, especially in fuel and ores and metals dependent economies. The dependent variable in Column 6 is the proxy for tariff evasion, as in Column 3. In contrast to Column 3, the interaction effect for the fuel export share is negative here, indicat- ing NTMs are not significantly associated with greater evasion in fuel exporting economies. 17 An explanation for this is that since NTMs provide legal asymmetric cost advantages (e.g., for firms with access to quotas), they make tariff evasion less attractive as a means to main- tain market share in fuel exporting economies. 5 Import monopolization and domestic prices A manifestation of the resource curse is higher prices in commodity exporting economies. Sachs and Warner (2001) showed that commodity export intensive economies had higher price levels relative to the global average in 1979. A question is whether import monopo- lization can account for this phenomenon. In theory, the relationship between concentration and prices is not obvious. If higher concentration is associated with higher fixed costs but lower marginal costs of importing, it could be associated with lower prices, even if also as- sociated with higher markups (e.g., as in a differentiated products Nash-in-prices game). Alternatively, higher concentration could be associated with higher prices, if the associated markups outweigh any marginal cost savings. To distinguish between these hypotheses, we relate import market concentration to ICP data on the average local currency domestic prices of 685 narrowly defined goods (i.e., par- boiled long-grain rice, basmati rice, brown rice—family pack). From this correlation, we infer whether, on average, marginal cost differences associated with import market concentration outweighs the markups associated with concentration in determining equilibrium prices. The spirit of this exercise is an accounting decomposition, not an attempt to estimate a causal relationship between prices and concentration. The industrial organization literature has long argued that such causal effect is not well-defined, because a variety of economic mech- anisms can cause a (positive or negative) correlation between prices and concentration (see Miller, Berry, Scott Morton, Baker, Bresnahan, Gaynor, Gilbert, Hay, Jin, Kobayashi et al., 2022). Our exercise is simply to estimate the correlation between prices and concentration, conditional on product-year fixed effects that account for global supply and demand. 18 LCU Define the price of product i in country c and year t as Pc,i,t . The ICP provides these prices for two years, 2011 and 2017. Formally, we estimate the regression LCU ln Pc,i,t = αi,t + β4 ln HHIc,s(i),t + β5 E LCU/U SD + β6 Xc,t + ϵc,i,t (3) where αi,t is a product-year fixed effect. Import market concentration HHIc,s(i),t is calculated at the level of the ICP basic heading s, one of 88 product groupings that include multiple products i. An example is the basic heading for rice, which includes 9 varieties, including parboiled long-grain rice, basmati rice, and brown rice—family pack. While at least one variety of rice is observed per country, not all varieties are observed in every country. Cal- culating the HHI at the basic heading level ensures the market definition is consistent and comparable across countries.10 Estimates of Equation 3 are reported in Table 4. In Column (1), the coefficient β4 = 0.07 (0.02) indicates a 1 percent increase in HHI (i.e., an increase of 42 units, since the mean value of HHI is 4,211 in Table 1, Panel A, Column 1) is associated with 0.07 percent higher prices. In Column (2), country-year controls are included and the coefficient falls to β4 = 0.04 (0.02), though it remains positive and statistically significant at the 5 percent level. The controls apparently explain some variation in prices correlated with HHI, but not all of it. The value added of these regressions, beyond the magnitude of the coefficient, is that they indicate that markups associated with higher concentration appear to outweigh any lower marginal costs of importing, at least on average. This is consistent with a model of a competitive world price and importers who are price-takers but have market power in local distribution. Importing firms in India, for example, have been shown to conform to this model (De Loecker, Goldberg, Khandelwal and Pavcnik, 2016). The association between import concentration and higher prices—a symptom of the resource curse—suggests 10 When experimenting with specifications we found the log of HHI to fit better than the level of HHI, consistent with the non-linear associations of HHI with prices, conditional on marginal cost, in the Cournot model. 19 import concentration is a novel channel through which the curse materializes. An exciting avenue for future research would exploit data on prices, domestic consumption, and market concentration to estimate demand and evaluate the consequences of import monopolization for consumer welfare. 6 Concluding remarks This paper identifies a novel channel for the ‘resource curse’: the monopolization of imports. Commodity export intensity is associated with more concentrated import markets, which can account for the higher price levels typically attributed to the Dutch disease. Trade policy measures and tariff evasion are mechanisms contributing to import market concentration. While economies’ export orientation has been the focus of the trade and development literature, the role of imports and import market structure has been overlooked. Openness to imports is generally thought to increase competition in an economy, but this effect could attenuate severely in the presence of importer market power. Further research could explore which domestic value chains can emerge from more competitive import markets, especially in commodity export intensive economies. Given the size of the import sector in developing countries, the policy agenda to de- monopolize imports is paramount. Direct policy interventions to increase import competi- tion include reforms that lower tariffs and remove NTMs such as quotas that restrict entry. The creation or strengthening of credible independent local bodies will help promote compe- tition in general and in the import sector in particular. Fighting anti-competitive practices can prevent the perpetuation of oligarchies who constitute an important lobby and can seize control of liberalization attempts. Another integral part of a competition agenda is trans- parency and data availability. In many countries, barriers to entry into the import sector like licenses and the exact role of state-owned enterprises are not easily measured. The flow of funds between public banks and other state-owned enterprises is opaque and leads to 20 asymmetric cost advantages. Transparent public procurement and the use of auctions to allocate import licenses could also increase import market competition. 21 Exhibits Figure 1: Import market concentration in natural resource-dependent countries and a potential mechanism (a) Import market concentration (b) Trade protection .15 GNB GAB 3000 LAO Tariff rate, applied, weighted mean, all products TLS Average HHI of country's import markets MAR BGD NPLBEN CMR GNB BDI STP SEN MWI ETH CIV GAB RWA TGO CPV 2000 MUS TGO MDG CPV MKD RWA .1 CMR STP BWA ZMB KEN SEN CRI PAKLKA JOR TZA EGY UGA SLV TZA ETH CIV DOM UGA GEO BDI COL MDG ALB MAR BEN KEN KHM GTM ECU URY DOM IND COM LKA JOR MWI KHM COM PER PRY COL .05 SRB ECU PRY URY HRV NPL EGY ZAF 1000 CHL PER CHL ZMB ZAF MEX MUS GEO GTM IDN MKD SRB SLV BGR ROU TLS BGR VNM ROU IDN 22 PAK ALB HRV CRI BGD VNM LAO BWA MEX IND 0 0 .2 .4 .6 .8 1 0 .2 .4 .6 .8 1 Commodity export share of total merchandise exports Commodity export share of total merchandise exports Notes: Panel A displays the average HHI of a country’s import markets and the country’s average commodity export share, where averages are taken over years. Within years, the HHI is the import value-weighted average of the log of the HHI across HS 6-digit products for a given country. The slope of the best fit line in Panel A is 0.86 (standard error = 0.19). Panel B displays the import value-weighted average tariff rate of a country and the country’s average commodity export share, where averages are taken over years. The slope of the best fit line in Panel B is 0.031 (0.022). Table 1: Import market concentration and commodity exports (1) (2) (3) (4) (5) (6) Dependent variable HHI HHI HHI HHI HHI HHI Panel A) All Products Commodity export share ∈ [0, 1] 901*** 1,238*** 438* 358** (132) (169) (228) (170) Export commodity price index (1=2012) 1,050* 767 (614) (488) Log GDP per capita -665*** -684*** (163) (161) Gini coefficient ∈ [0, 1] -2,904*** -1,645*** (574) (553) Control of corruption -484*** -553*** (65) (64) Regulatory quality -391*** -291*** (74) (74) R-squared 0.00 0.40 0.75 0.76 0.75 0.76 Number of observations 2,190,652 2,188,974 2,176,363 1,716,738 2,249,099 1,789,513 Dependent variable mean 4,219 4,216 4,198 4,121 4,184 4,107 Country-product fixed effects No No Yes Yes Yes Yes Product-year fixed effects No Yes Yes Yes Yes Yes Panel B) By Product Category Broad Economic Category (BEC) Primary Capital Consu- Primary Capital Consu- mption mption Commodity export share [0,1] 500** 489* 458* (219) (254) (264) Export commodity price index (1=2012) 1,141* 338 1,616** (581) (769) (686) R-squared 0.72 0.73 0.72 0.72 0.73 0.72 Number of observations 103,755 334,895 613,310 107,580 345,683 632,973 Dependent variable mean 6,015 3,544 3,630 5,990 3,529 3,622 Country-product fixed effects Yes Yes Yes Yes Yes Yes Product-year fixed effects Yes Yes Yes Yes Yes Yes Notes: The dependent variable is the HHI, i.e., the sum of squared percent import market shares in a country-product-year, which has a maximum value of 10,000. Products are classified by the 6 digit Harmonized System. The number of observations summed across columns (1)-(3) or (4)-(6) in Panel B is smaller than the total number of import markets in columns (3) and (5), respectively, of Panel A because intermediates in the BEC product categories are not considered in Panel B. Standard errors clustered at the country-year level are in parentheses. ***, **, and * indicate significance at 1%, 5%, and 10% levels, respectively. 23 Table 2: Alternative measures of import market concentration (1) (2) (3) (4) (5) (6) Dependent variable Share of Share of Log number Log number Log Share of the largest the 4 largest of of source unit commodity importer importers importers countries value exporters Panel A) Commodity export share ∈ [0, 1] 0.04* 0.02 -0.25* -0.23*** 0.33* 0.00 (0.02) (0.01) (0.13) (0.08) (0.18) (0.02) R-squared 0.73 0.85 0.94 0.90 0.82 0.56 Number of observations 2,176,363 2,176,363 2,176,363 2,176,363 1,818,137 1,969,637 Dependent variable mean 0.530 0.807 2.616 1.762 1.739 0.189 Panel B) Export commodity price index (1=2012) 0.09* 0.04 -0.40 -0.58*** 0.95*** 0.01 (0.05) (0.04) (0.35) (0.20) (0.27) (0.06) Number of observations 2,249,099 2,249,099 2,249,099 2,249,099 1,883,317 2,034,411 R-squared 0.73 0.85 0.94 0.90 0.82 0.57 Dependent variable mean 0.53 0.81 2.63 1.77 1.74 0.19 Country-product fixed effects Yes Yes Yes Yes Yes Yes Product-year fixed effects Yes Yes Yes Yes Yes Yes Notes: Dependent variables in columns 1, 2, and 6 range from 0 to 1. Standard errors clustered at the country-year level are in parentheses. ***, **, and * indicate significance at 1%, 5%, and 10% levels, respectively. 24 Table 3: Trade policy mechanisms for import market concentration (1) (2) (3) (4) (5) (6) Dependent variable HHI HHI Evasion HHI HHI Evasion ∈ [−1, 1] ∈ [−1, 1] Tariff ∈ [0, 1]] 3,514*** 1,794* -0.48*** (1,019) (1,073) (0.13) Commodity export share x Tariff -1,636 (1,399) Fuel export share x Tariff 8,696*** 0.36** (1,907) (0.17) Ores and metals export share x Tariff 1,459 -0.24 (1,827) (0.18) Food export share x Tariff -9,887*** -0.93*** (1,551) (0.16) NTM ∈ {0, 1} 35 1 -0.00 (40) (47) (0.01) Commodity export share x NTM 58 (137) Fuel export share x NTM 146 -0.02 (186) (0.02) Ores and metals export share x NTM 532*** 0.03 (201) (0.03) Food export share x NTM -126 -0.04** (174) (0.02) Number of observations 1,848,851 1,848,851 1,848,851 342,148 342,148 342,148 R-squared 0.47 0.47 0.19 0.52 0.52 0.17 Dependent variable mean 3,896 3,896 0.0439 3,429 3,429 0.03 Country-year fixed effects Yes Yes Yes No Yes Yes Product-year fixed effects Yes Yes Yes No No No Notes: Commodity, fuel, ores and metals, and food exports shares are measured in difference from their median value. Non- tariff measures (NTMs) are those those in Chapters E and F of the UNCTAD classification. Standard errors clustered at the country-year and country-product levels are in parentheses. ***, **, and * indicate significance at 1%, 5%, and 10% levels, respectively. 25 Table 4: Import market concentration and domestic prices (1) (2) Dependent variable Log Log domestic domestic price price Log HHI 0.07*** 0.04** (0.02) (0.02) Log nominal exchange rate with US dollar 0.92*** 0.87*** (0.01) (0.01) Log GDP per capita -0.04 (0.02) Gini coefficient 5.34*** (0.31) Control of Corruption 0.14*** (0.02) Regulatory Quality -0.23*** (0.03) R-squared 0.91 0.90 Number of observations 19,215 15,900 Dependent variable mean 6.53 6.80 Product-year fixed effects Yes Yes Notes: The HHI is calculated at the level of the ICP basic heading level (e.g., rice) which is constructed to include at least one product in each country and maps roughly one-to-one to 2 digit HS codes. There are 88 distinct basic headings. Domestic prices are observed for 685 distinct goods (e.g., long grain rice; basmati rice) though not all are observed in each country. Standard errors clustered at the ICP basic heading-year are in parentheses. ***, **, and * indicate significance at 1%, 5%, and 10% levels, respectively. 26 References Aghion, Philippe, Nick Bloom, Richard Blundell, Rachel Griffith, and Peter Howitt, “Competition and innovation: An inverted-U relationship,” The Quarterly Jour- nal of Economics, 2005, 120 (2), 701–728. as, Pol, Teresa C Fort, and Felix Tintelnot, “The margins of global sourcing: Antr` Theory and evidence from US firms,” American Economic Review, 2017, 107 (9), 2514– 2564. Arezki, Rabah, Valerie A Ramey, and Liugang Sheng, “News shocks in open economies: Evidence from giant oil discoveries,” The Quarterly Journal of Economics, 2017, 132 (1), 103–155. Benkard, C Lanier, Ali Yurukoglu, and Anthony Lee Zhang, “Concentration in product markets,” Working Paper 28745, National Bureau of Economic Research 2021. Bhagwati, Jagdish, “On the underinvoicing of imports,” Bulletin of the Oxford University Institute of Economics & Statistics, 1964, 27 (4), 389–397. es, Quantitative techniques for competition and antitrust Davis, Peter and Eliana Garc´ analysis, Princeton University Press, 2009. der Ploeg, Frederick Van, “Natural resources: curse or blessing?,” Journal of Economic Literature, 2011, 49 (2), 366–420. Edmond, Chris, Virgiliu Midrigan, and Daniel Yi Xu, “How costly are markups?,” Journal of Political Economy, 2023, 131 (7), 1619–1675. Fernandes, Ana M, Caroline Freund, and Martha Denisse Pierola, “Exporter be- havior, country size and stage of development: Evidence from the exporter dynamics database,” Journal of Development Economics, 2016, 119, 121–137. 27 Fisman, Raymond and Shang-Jin Wei, “Tax rates and tax evasion: evidence from “missing imports” in China,” Journal of Political Economy, 2004, 112 (2), 471–496. Frankel, Jeffrey A, “The natural resource curse: A survey of diagnoses and some prescrip- tions,” Commodity price volatility and inclusive growth in low-income countries, 2012, pp. 7–34. Freund, Caroline, Rich People Poor Countries: The Rise of Emerging-Market Tycoons and their Mega Firms, Peterson Institute for International Economics, 2016. and Martha Denisse Pierola, “Export superstars,” Review of Economics and Statis- tics, 2015, 97 (5), 1023–1032. Gaubert, Cecile, Oleg Itskhoki, and Maximilian Vogler, “Government policies in a granular global economy,” Journal of Monetary Economics, 2021, 121, 95–112. Goldberg, Pinelopi Koujianou and Tristan Reed, “Is the global economy deglobal- izing? If so, why? And what is next?,” Brookings Papers on Economic Activity, March 2023. Gruss, Bertrand and Suhaib Kebhaj, Commodity terms of trade: A new database, International Monetary Fund, 2019. Harding, Torfinn and Anthony J Venables, “The implications of natural resource exports for nonresource trade,” IMF Economic Review, 2016, 64 (2), 268–302. Isham, Jonathan, Michael Woolcock, Lant Pritchett, and Gwen Busby, “The vari- eties of resource experience: natural resource export structures and the political economy of economic growth,” The World Bank Economic Review, 2005, 19 (2), 141–174. Javorcik, Beata S and Gaia Narciso, “WTO accession and tariff evasion,” Journal of Development Economics, 2017, 125, 59–71. 28 Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi, Worldwide Governance Indicators, World Bank, 2010. Krueger, Anne O, “The political economy of the rent-seeking society,” The American Economic Review, 1974, 64 (3), 291–303. Leone, Fabrizio, Rocco Macchiavello, and Tristan Reed, “The High and Falling Price of Cement in Africa,” American Economic Journal: Applied Economics, 2024. Loecker, Jan De, Pinelopi K Goldberg, Amit K Khandelwal, and Nina Pavcnik, “Prices, markups, and trade reform,” Econometrica, 2016, 84 (2), 445–510. Miller, Nathan, Steven Berry, Fiona Scott Morton, Jonathan Baker, Timothy Bresnahan, Martin Gaynor, Richard Gilbert, George Hay, Ginger Jin, Bruce Kobayashi et al., “On the misuse of regressions of price on the HHI in merger review,” Journal of Antitrust Enforcement, 2022, 10 (2), 248–259. Mishra, Prachi, Arvind Subramanian, and Petia Topalova, “Tariffs, enforcement, and customs evasion: Evidence from India,” Journal of Public Economics, 2008, 92 (10- 11), 1907–1925. Nocke, Volker and Michael D Whinston, “Concentration Thresholds for Horizontal Mergers,” American Economic Review, 2022, 112 (6), 1915–48. Rigterink, Anouk, Tarek Ghani, Juan Lozano, and Jacob Shapiro, “Mining Com- petition and Violent Conflict in Africa: Pitting Against Each Other,” University of Wash- ington mimeo, 2022. Robinson, James A, Ragnar Torvik, and Thierry Verdier, “Political foundations of the resource curse: A simplification and a comment,” Journal of Development Economics, 2014, 106, 194–198. Ross, Michael L, “The oil curse,” in “The Oil Curse,” Princeton University Press, 2012. 29 Sachs, Jeffrey D. and Andrew M. Warner, “The curse of natural resources,” European Economic Review, 2001, 45 (4-6), 827–838. Tornell, Aaron and Philip R Lane, “The voracity effect,” American Economic Review, 1999, 89 (1), 22–46. Tullock, Gordon, “The welfare costs of tariffs, monopolies, and theft,” Economic inquiry, 1967, 5 (3), 224–232. UNCTAD, “International Classification of Non-tariff Measures,” Technical Report 2019. , “Key Statistics and Trends in International Trade 2020,” 2021. UNCTAD, “TRAINS,” https://trainsonline.unctad.org/home 2024. Accessed Febru- ary 2024. United States Federal Trade Commission and Department of Justice, “Merger Guidelines,” 2023. Venables, Anthony J, “Using natural resources for development: why has it proven so difficult?,” Journal of Economic Perspectives, 2016, 30 (1), 161–84. World Bank, “Average Prices from the International Comparison Program.,” https://thedocs.worldbank.org/en/doc/550841576791478325-0050022019/ original/ICPDataAccessandArchivePolicyDecember2019.pdf 2011, 2017. , “World Development Indicators,” https://datatopics.worldbank.org/ world-development-indicators/ 2024. Accessed June 20, 2024. , “World Integrated Trade System,” 2024. Accessed February 12, 2024. 30 Online Supplement to an, Nguyen, and Reed Arezki, Fernandes, Merch´ “Natural Resource Dependence and Monopolized Imports” 31 Table S1: Import market structure and commodity exports by country No. Country Min. Max. Median Average Commodity Fuel Ores & metal Food year year HHI HHI export export export export share share share share 1 Albania 2007 2021 3285 4258 0.27 0.11 0.09 0.07 2 Bangladesh 2005 2015 2885 3983 0.06 0.01 0.00 0.04 3 Benin 2016 2021 5632 5949 0.30 0.00 0.01 0.30 4 Botswana 2004 2010 3321 4201 0.20 0.00 0.16 0.03 5 Bulgaria 2001 2006 2555 3531 0.32 0.09 0.13 0.10 6 Burundi 2010 2021 6408 6413 0.59 0.00 0.06 0.53 7 Cabo Verde 2010 2021 4069 4914 0.84 0.00 0.00 0.84 8 Cambodia 2016 2021 4489 5134 0.05 0.00 0.00 0.05 9 Cameroon 2007 2017 4019 4785 0.67 0.34 0.06 0.27 10 Chile 1997 2021 2357 3403 0.77 0.01 0.53 0.22 11 Colombia 1997 2021 2468 3465 0.65 0.47 0.01 0.17 12 Comoros 2016 2021 5317 5730 0.65 0.00 0.01 0.64 13 Costa Rica 2010 2021 2922 3926 0.40 0.00 0.01 0.38 14 Cote d’Ivoire 2000 2021 4212 4915 0.69 0.20 0.00 0.49 15 Croatia 2007 2015 2167 3257 0.29 0.13 0.04 0.12 16 Dominican Republic 2002 2021 3081 4052 0.26 0.04 0.02 0.20 17 Ecuador 2002 2021 2739 3791 0.87 0.48 0.01 0.38 18 Egypt, Arab Republic 2005 2016 2442 3495 0.52 0.33 0.05 0.14 19 El Salvador 2006 2021 3304 4252 0.24 0.03 0.01 0.20 20 Ethiopia 2012 2021 3756 4605 0.79 0.00 0.00 0.79 21 Gabon 2009 2021 4471 5134 0.65 0.62 0.03 0.00 22 Georgia 2000 2021 3689 4582 0.55 0.04 0.21 0.29 23 Guatemala 2005 2013 3082 4070 0.52 0.05 0.05 0.41 24 Guinea-Bissau 2014 2018 8068 7059 0.91 0.00 0.00 0.91 25 India 2016 2021 1527 2516 0.28 0.13 0.04 0.11 26 Indonesia 2020 2020 1978 2929 0.45 0.16 0.06 0.24 27 Jordan 2008 2021 3837 4688 0.26 0.01 0.09 0.17 28 Kenya 2006 2021 3403 4310 0.54 0.04 0.04 0.46 29 Lao PDR 2015 2021 5641 5998 0.64 0.17 0.25 0.21 30 Madagascar 2007 2021 4578 5197 0.58 0.04 0.20 0.34 31 Malawi 2005 2021 4767 5321 0.87 0.00 0.03 0.84 32 Mauritius 2000 2021 3554 4431 0.34 0.00 0.01 0.33 33 Mexico 2011 2021 1598 2539 0.18 0.08 0.03 0.07 34 Morocco 2002 2013 2417 3504 0.34 0.04 0.10 0.20 35 Nepal 2011 2014 2886 3898 0.26 0.00 0.04 0.22 36 North Macedonia 2008 2018 3213 4187 0.23 0.04 0.06 0.13 37 Pakistan 2019 2021 2410 3623 0.24 0.01 0.04 0.19 38 Paraguay 2000 2021 3338 4328 0.88 0.35 0.01 0.53 39 Peru 2000 2021 2568 3567 0.68 0.08 0.41 0.19 40 Romania 2005 2011 1864 2806 0.18 0.08 0.04 0.06 41 Rwanda 2002 2016 5198 5582 0.82 0.00 0.36 0.46 42 Sao Tome and Principe 2017 2019 5166 5714 0.84 0.00 0.02 0.82 43 Senegal 2000 2020 4319 5026 0.55 0.17 0.05 0.33 44 Serbia 2006 2007 2377 3403 0.33 0.03 0.11 0.19 45 South Africa 2010 2021 2108 3058 0.48 0.10 0.27 0.10 46 Sri Lanka 2016 2021 3107 4040 0.28 0.02 0.01 0.25 47 Tanzania 2003 2021 3608 4477 0.43 0.01 0.10 0.32 48 Timor-Leste 2018 2021 4757 5357 0.97 0.67 0.00 0.30 49 Togo 2015 2021 5824 6157 0.41 0.06 0.13 0.22 50 Uganda 2011 2020 3669 4536 0.59 0.03 0.01 0.55 51 Uruguay 2001 2021 3118 4105 0.62 0.02 0.00 0.59 52 Vietnam 2018 2021 1569 2577 0.12 0.01 0.01 0.09 53 Zambia 2010 2021 3567 4451 0.85 0.01 0.76 0.08 32 Table S2: Summary statistics of independent variables and controls Variable Mean Std. Dev. Commodity export share .53 .24 Food export share .31 .22 Fuel export share .12 .18 Ores and metals export share .11 .17 Commodity price index (1=2012) .99 .03 Log GDP per capita 8.82 .85 Gini coefficient .45 .08 Control of corruption -.32 .67 Regulatory quality -.12 .61 Tariff, applied .08 .05 NTM, Chapter E or F .44 .50 Evasion gap .05 .51 Notes: Summary statistics are based on the estimating sample for all products used in Table 1. 33 Table S3: Natural resource dependence and control of corruption (1) (2) (3) Dependent variable Control of Control of Control of corruption corruption corruption Commodity export share ∈ [0, 1] -0.39** (0.17) Fuel export share ∈ [0, 1] -0.64*** (0.14) Ores and metals export share ∈ [0, 1] -0.27 (0.26) Food export share ∈ [0, 1] -0.02 (0.23) Export commodity price index (1=2012) -0.59 (0.48) Log GDP per capita 0.58*** 0.65*** 0.59*** (0.11) (0.11) (0.11) Gini coefficient 0.90* 1.18** 0.55 (0.51) (0.52) (0.43) Regulatory quality 0.52*** 0.49*** 0.48*** (0.06) (0.06) (0.06) R-squared 0.95 0.95 0.95 Number of observations 488 488 508 Dependent variable mean -0.301 -0.301 -0.313 Country fixed effects Yes Yes Yes Year fixed effects Yes Yes Yes Note: Robust standard errors are in parentheses. ***, **, and * indicate significance at 1%, 5%, and 10% levels, respectively. 34