THE ROLE OF NATIONALISM IN THE UNITED STATES ECONOMY by MELISSA LIU A THESIS Presented to the Department of Economics and the Robert D. Clark Honors College in partial fulfillment of the requirements for the degree of Bachelor of Science June 2019 An Abstract of the Thesis of Melissa Liu for the degree of Bachelor of Arts in the Department of Economics to be taken June 2019 Title: The Role of Nationalism in the United States Economy Approved: _______________________________________ Woan Foong Wong Do instances of economic nationalism translate into consumer behavior concerning American import levels? Individual consumer biases and economic nationalism have been commonly researched in marketing and sociology, but have rarely been translated into a larger macroeconomic scale. In April 2017, President of the United States, Donald Trump, signed an executive order, “Buy American and Hire American.” In doing so, he called upon the “buy domestic” rhetoric that has been echoed through centuries of American politics. This paper will analyze the effects of this decision through a difference in differences model used in Mitchell Morey’s 2015 paper on home bias in trade. Results find that “Buy American and Hire American” caused imports to decrease for goods covered by the policy while causing overall imports to increase. This paper looks at the consumer implications of domestic content requirements, underscoring the repercussions of such politically attractive policies. ii Acknowledgements I express my deepest gratitude to Professor Woan Foong Wong, for guiding me through this process while learning about it at the same time, for encouraging me through rough patches, and for reading through this thesis more than anyone else. Your passion and nuanced understanding of the field has made this thesis process exciting statistical and narrative challenge. I also would like to thank Professor Louise Bishop and Professor Tim Duy, the only two professors I willingly enrolled for additional classes with after the first, because their zeal and drive in their respective fields translated so poignantly through their teaching. I am also grateful for my friends for supporting me through solidarity, hugs, laughter, and many sugary snacks. Last, but not least, I express my greatest thanks to my family, for supporting me for so many years, through failures and successes, through tears of laughter and sadness, to get me to where I am today. This thesis would not have been finished without any one of you. iii Table of Contents Introduction 1 Demand & Utility 2 Buy American Act 3 Literature Review 5 Economic Nationalism 5 Country-of-Origin Effect and Home Bias 7 Protectionism in Practice 13 Research Question 21 Hypothesis 21 Methodology 22 Model 22 Proxy of Economic Nationalism 23 Price Effects 24 Simplicity 24 Chosen Goods 26 Model Specifications 27 Software 28 Results 29 Heteroskedasticity and Multicollinearity 32 Robustness Checks 35 Discussion 37 Model Limitations 40 Conclusion 42 Appendix A: STATA Code 44 Appendix B: Preliminary Regression Results 47 Appendix C: Commodities 48 Bibliography 55 iv List of Figures Figure 1: United States Iron & Steel and Aluminum Annual Imports 25 Figure 2: White’s Test for heteroskedasticity 32 Figure 3: United States Public and Private Construction Spending 39 Figure 4: United States Infrastructure Spending 40 List of Tables Table 1: Observation Summary Statistics ...................................................................... 27 Table 2: Difference in differences regression results (not robust) ................................. 30 Table 3: Multicollinearity between variables ................................................................. 32 Table 4: Robust OLS regression results with p-values .................................................. 33 Table 5: Robust OLS regression results with standard errors ........................................ 34 Table 6: Final OLS regression results ............................................................................ 35 Table 7: Robustness checks, p-values ............................................................................ 36 Table 8: Robustness checks, standard errors .................................................................. 37 v Introduction In 2013, Walmart launched an initiative committed to selling and supporting U.S.-made products and U.S. manufacturers. According to their corporate website, the company’s $250 billion investment was estimated to create one million new American jobs over the ten-year initiative. Similarly, in his successful campaign for the United States presidency, Donald Trump repeatedly boasted about his plan to bring jobs back to the U.S., slapping tariffs on German made cars and restricting the imports of Chinese-produced technology. The idea of promoting locally or nationally produced goods is not a phenomenon unique to the twenty first century nor to the United States. Even at the very inception of the United States, the Boston Tea Party demonstrated against British imports (Frank 1999). This type of rhetoric goes hand in hand with protectionist policies: tariffs, quotas, and other ways governments and consumers isolate their domestic economy from international economy. Protectionist policies ultimately bring an economy further away from free trade – international trade without restrictions – and closer to autarky – complete economic independence. It is generally accepted by economists that under free trade, economies are pushed to greater efficiencies and higher states of general welfare, and thus bjnkgive consumers the benefit of more choice. Conversely, protectionism has been shown to hurt both the nations attempting to protect their industries and those nations its being protected against (Larch & Lechthaler 2009). It begs the question, do American consumers care if their products are made in the U.S.? Is this observable enough to affect levels of imports? My research will investigate if instances of increased “Made in America” national rhetoric have a measurable macroeconomic effect. To answer this question, I will first contextualize my research with economic and historical background; then with the support of past research, argue against the implementations of protectionist policies in large and developed economies such as the United States; present my own findings on the negative effect of a specific American policy, The Buy American Act of 1933; and finally conclude with the debated role of economic nationalism in the American economy. Demand & Utility Economists generally consider the studies of employment and costs of production the “supply side” of economics, while the studies of consumer behavior and preferences are considered the “demand side.” Although it may be clear why Americans would support initiatives that promote domestic employment, it is not as salient if such initiatives would translate into American consumer preferences. At the most rudimentary level, consumer behavior can be explained by the Law of Demand, which states the inverse relationship between quantity demanded and price. Although the model does not account for many other variables affecting consumer behavior, it is commonly accepted that the Law of Demand holds for most circumstances. Another large driving economic principle is utility. Given prices and a budget constraint, consumers will buy what gives them the most utility. How much utility, or units of “utils”, each consumer gains from a product is a function of their preferences. Analogously, a country can be thought of as one large consumer that 2 imports goods that its preferences demand. My research looks at these preferences for U.S. produced goods through ebbs and flows of national trade data. Buy American Act Passed in 1933 on the last day of the Hoover administration, the Buy American Act (BAA) is a domestic content requirement policy that aims to focus production and acquisition of goods within the United States, stating: [o]nly unmanufactured articles, materials, and supplies that have been mined or produced in the United States, and only manufactured articles, materials, and supplies that have been manufactured in the United States substantially all from articles, materials, or supplies mined, produced, or manufactured in the United States, shall be purchased for public use. The Act focuses on the procurement of construction materials for public projects. The Federal Acquisition Regulation (FAR) regulates and enforces the procurement of materials for public projects. Price preferences are artificially controlled through adding a certain percentage of the lowest offer price plus duties. In general, 6% is added to the lowest foreign offer when the lowest domestic offer is from a large business, 12% is added to the lowest foreign offer when the lowest domestic offer is from a small business, and up to 50% for Department of Defense procurements. After these rate changes, if the domestic producer is tied for the lowest offer, the agency must go with the domestic producer. If after these applications, the foreign offer is still the lowest, choosing the foreign products is permitted due to an “unreasonable” additional cost from choosing the domestic products. Beyond when the price inflation still favor the foreign imports, the President of the United States can waive BAA at their discretion. Buy American is generally waived due to public interest, non-availability of supplies, if 3 the end-product of the materials is for commercialized purposes, and to the aforementioned unreasonable costs to the taxpayer. How much does public procurement matter in the scope of the American economy? Annually, public procurement in the United States amounts to about $1.9 trillion, which translates to approximately ten percent of annual Gross Domestic Product (GDP), the total value of goods and services produced in a country. Import content, alone, makes up $95 billion (Hufbauer & Jung 2017). Thus, policies like ‘Buy American’ not only have lasting and compounding effects on the American construction industry, but on the entire economy. 4 Literature Review Economic Nationalism To work within the critical lens of economic nationalism, it is prudent to evaluate the individual, societal, and psychological motivations of economic nationalism. Much of the research done to define the term, “economic nationalism,” comes from psychological and sociological sources. In the 1940s, Hans Kohn defined nationalism to be the “state of mind in which the individual identified himself with the ‘we-group’ to which supreme loyalty is given” (Kohn, 1944; Baughn and Yaprak, 1996). Groups begin linking distinctions with nationalistic orientation, paving a way for prejudice and discrimination when the ‘in- group’ perceives physical, social or economic threats from the ‘out-group’ (Baughn and Yaprak, 1996). It must be noted, however, that nationalism differs from ‘patriotism’, where a distinction and positive image of the ‘in-group’ does not necessarily suggest a negative image and dislike of the ‘out-group.’ Economic nationalism, then, can be interpreted as protective from those perceived threats. Mostly commonly, economic nationalism suggests that the economy is ‘zero-sum’, where economic gains that are not of the ‘in-group’ are considered losses. It is most clearly observed through policy interventions that favor one’s own nation. Free trade, as discussed earlier, facilitates international exchanges of production; economic nationalism seeks to alter the direction or “nature of these trade, capital, and technology flows” (Baughn and Yaprak, 1996). Primary methods include tariffs, a tax on imports or exports; quotas, limitations on imports; restrain agreements, and duties. These terms generally carry negative and 5 restrictive economic connotations, but economic nationalism can also come in more ‘benevolent’ forms such as domestic subsidies or increasing regulatory standards. While these interventions have clear quantitative effects on trade, instances of economic nationalism carry through international relationships that play large roles in determining trade agreements. The role of these trade relationships are magnified in the world of globalization, as the international connections in private firms give way to increased opportunity for friction and conflict (Baughn and Yaprak, 1996). Furthermore, the economic nationalism at the policy level can permeate and be expressed at through consumer behavior. Johnson (1967) and Macesich (1985) classify economic nationalism as a form of taste and economic discrimination, linking it to Becker’s (1957) work on racism. Becker suggests that psychologically, individuals can have a “taste for discrimination,” where they choose the psychological satisfaction of avoiding and discriminating a group over potential material gain, such as price (Becker, 1967). This thesis examines if there exists a taste for discrimination that can be observed at the macroeconomic level. To observe any effect, this taste for discrimination must be quantified in a measurable way. Only then can it be argued as an instance of nationalism, which I will later describe as a “treatment” of certain tradable goods. Past studies have indicated that nationalism is often positively correlated with ethnocentrism, authoritarianism, and conservatism while negatively correlated with internationalism (Eckhardt, 1991; Sampson & Smith, 1957). Research has shown that ethnocentrism correlates to resistance to the immigration of foreign workers and foreign investment (Johnson, 1965). Additionally, consumer ethnocentrism can be a result of economic competition 6 (Ray, 1984) that can be seen through consumer bias against foreign produced goods. These biases can be expressed through country-of-origin effects and home bias. Country-of-Origin Effect and Home Bias Given two options of the same good, there will be factors that pull a consumer to one good over the other. A product’s country-of-origin, for example, plays a large role in consumer habits. It is commonly accepted that consumers do not view goods produced by different countries as homogenous. Country-of-origin effect generalizes the role of where a product is produced, branded, and “made” on consumer behavior. Over 700 studies have established and verified the existence and degree of impact of the country-of-origin effect (Papadopoulos and Heslop, 2002). The effect can be broken down with several classes of goods, types of consumers, and different countries as producers. Costa et al. (2016) utilized a multidimensional structure of country-of-origin image in order to gauge a more accurate generalization of the impact. The study found that for different types of goods, different emotional and logical responses were associated with different countries of origin. Differentiated responses likely go hand in hand with well-known consumer biases that are a result of history, popular culture, and the media, like German cars or Korean refrigerators (Hamzaoui-Essoussi et al., 2011). On the other hand, consumers are not always working with complete and decision- altering information. Chao (1998) found that consumers do no significantly differentiate the perceived quality of goods when given knowledge of country-of-assembly, country- of-design, and parts-source country. Country-of-origin studies lay important groundwork regarding general consumer habits for my research; they allow me, and 7 other researchers, to treat country-of-origin as better defined variables when developing further research aims. It also must be noted that consumer preferences are neither uniform nor do they conform perfectly to theoretical models. This could potentially explain why in practice, the world trades 50% less than the neoclassical Heckscher-Ohlin-Vanek trade model predicts (Trefler, 1995). In a study of the causes and consequences of regional taste differences, Atkin (2013) finds that taste differences that arose from accessible and abundant foods from prior generations form regional purchasing habits that can be at the cost of price and taste. Such studies underscore how, outside of traditional marketing, cultural and societal habits can explain why consumers, groups, and countries make decisions that favor domestic production that sacrifices efficiency. Farmers in western Kenya, for example, can set a decent premium on their domestically grown maize relative to other available maize because consumers are more assured of its safety and free of aflatoxin (Hoffman and Gatobu 2012). Their research shows how a cultural food safety concern can translate into higher than expected levels of domestic agricultural production. In line with my research interests, it has also been shown that national interests and concerns can translate into consumer behavior. Morey (2016) research finds that on average, the Malagasy population in Madagascar’s capital were willing to pay a 9% price premium on labeled Malagasy produced rice over labeled French produced rice, despite preferring the French rice when nothing was labeled. Madagascar’s colonial relationship with France, however, played a significant role in the price a participant was willing to pay – Malagasies with surveyed anti-French sentiment were willing to 8 pay a statistically significant more amount for Malagasy produced rice than those with neutral or pro-France sentiments. Studies on ‘Buy Local’ movements have also indicated that consumers are willing to buy locally grown produce with a higher price tag for the sake of environmental sustainability or supporting local businesses (Darby et al. (2006); Hu et al. (2012)). These studies, like mine intends, test consumers’ willingness to pay more for the sake of general welfare. However, these studies only imply macroeconomic effects – Morey (2016), for example, suggests that the 9% premium Malagasy consumers are willing to pay could mean about a 5% decrease in French rice imports. Drawing heavily from Morey’s models, I will look directly at U.S. imports during a resurgence of “Made in USA” rhetoric to see if the suggested consumer preference for domestically produced goods is translate to the import and export levels. Morey’s work is an excellent example of how discrimination against foreign products is only one of the components of economic nationalism. In comparison to other potential explanations of home bias, a consumer’s perception of “domestic” versus “foreign” producers define economic nationalism (Reich, 1991). Baughn and Yaprak (1996) and Kahan (1967) synthesize it from the American consumer perspective: Economic nationalism is based nationalism is based on two propositions that, for brevity, may be stated as follows: 1. American Motors is better for the United States than British Motors and therefore deserves support. 2. What is good for the United States is good for American Motors. 9 Reich (1991) argues that economic nationalism reflects the belief that the success of domestic firms also ensures the well-being of a nation’s citizens, where domestic firms serve as the “intermediary” between the national economy and individual welfare. This type of rhetoric is used to justify protection of domestic industries and firms against their foreign counterparts. The economy is thus reduced to zero-sum, where investment and consumption of foreign goods is interpreted as “betrayals” or losses to the domestic economy and thus, the nation’s citizens. Baughn and Yaprak (1996) look at Realistic Group Conflict Theory to map out the psychological dynamic underpinning economic nationalism. Utilizing the “in- group” and “out-group” popularized by political scientists and sociologists, the authors argue that in the scope of the economy, domestic companies become the “in- companies” while foreign firms represent the “out-companies”. Through two surveys of undergraduate students, the authors focused on five measurements: economic nationalism, nationalism, patriotism, internationalism, and economic threat and job insecurity. In particular, restrictions on foreign firms, restrictions on foreign investment, restriction on immigration of workers, formal barriers to foreign products, “buy domestic” sentiment, intellectual property, domestic production by American firms, and general “U.S. first” orientation in regard to trade and competition, were used as metrics to determine economic nationalism (Baughn and Yaprak 1996). They used measurements based on the work of Kosterman and Feshbach (1989) for nationalism, patriotism, and internationalism. Economic threat and job insecurity were measured using Sharma et al.’s (1992) criteria. Results found that economic nationalism was predicted by perception of economic threat posed by foreign competitors (r = .52, p 10 .01). The survey also indication that despite substantial relationship between measures for nationalism and economic nationalism, patriotism showed to be a predictor for general nationalism but not for economic nationalism. Above all, the authors found that perceived economic threat was the strongest predictor for nationalism (p < 0.01). Their data suggests that “readiness” to support economic nationalism is a function of the perceived economic threat posed by foreign industries and competition (Baughn and Yaprak 1996). This perceived threat characterizes the context surrounding the passage of The Buy American Act of 1933. In the throes of the Great Depression, the United States was experiencing an annual unemployment peak of 23.6% and Gross Domestic Product (GDP) fall of 12.9% (United States Census Bureau). Periodic strengthening or reinforcements of Buy American can be similarly contextualized: Buy American provisions were included in President Barack Obama’s American Recovery and Reinvestment Act of 2009 following the 2008 stock market crash and again included as a part of President Donald Trump’s election promises to bring manufacturing jobs back to the United States. Gary Clyde Hufbauer and Eujin Jung from the Peterson Institute for International Economics synthesize this phenomenon, stating that “Buy American provisions are often enacted because politicians associate the patriotic slogan with the creation of domestic jobs” (2017). Reinforcements ‘Buy American’ policies are strong examples of economic nationalism in the United States. In her book, “Buy American: the Untold Story of Economic Nationalism,” Dana Frank draws a thorough narrative of economic nationalism over the course of American history (1999). Through the American Revolution, the Great Depression, and the period 11 of globalization, Frank especially scrutinizes the role of corporations in capitalizing on the working classes’ fears. She states, “Buy American campaigns blind us to corporate capital’s willingness to simultaneously play the nationalist card at home and flee overseas with its investment dollars” (251) From one of the first American demonstrations of economic nationalism, the Boston Tea Party, were cries from the working class opposing infringement on their economic democracy. Frank contends that self-interested corporations have perverted those original cries to consolidate capital. In particular, Frank argues that through and since the twentieth century, economic nationalism in the United States has become synonymous with anti Asian-American racism and Yellow Perilism (251). The “perceived threats” that Baughn and Yaprak attribute as a primary motivator for economic nationalism, is argued to be largely carried through American characterizations of Asian Americans as “sneaky, dangerous, inappropriate trading partners” (251). Frank’s narrative extends seamlessly into the twenty-first century, where the fast-growing Chinese economy has become a popularized enemy of American industries and jobs. Thus, it only makes sense that yet again, ‘Buy American’ has become a popular rhetorical device of the U.S. government. Frank’s work does an excellent job tying in Baughn and Yaprak’s findings (1996) to the United States. While research relating to the of fields marketing, sociology, and psychology looks heavily into the consumer implications, research generally (and very reasonably) does not consider macroeconomic data. Thus, a better framing of economic nationalism must require a deeper look into empirical analyses of protectionist policies. 12 Protectionism in Practice Only three years preceding the Buy American Act, Congress passed perhaps one of the most infamous American protectionist policies, the Smoot-Hawley Tariff Act of 1930. Popularly blamed for the rise of international protectionism (including by the League of Nations), slashing U.S. trade activity, and plunging the then-fragile American economy deeper into recession and depression, the Smoot-Hawley Tariff Act has had its fair share of analysis (Irwin 1996). Douglas A. Irwin looks into the 40% fall of U.S. imports over the two years following the passage of Smoot-Hawley. He found that the 20% average of tariff increases translated approximately into a 5-6% increase in the relative price of imports which further suggested a 4-8% fall in import volume (1996). Exacerbated by the rampant deflation in the early 1930s, his estimates found that the effective tariff amounted about 30% more, causing up to a 20% decrease in imports. Irwin’s work emphasizes the multiplicative role that rises in prices can have on imports. Furthermore, international retaliation and responsive protectionist policies against the United States only augmented these economic consequences. Intended to protect American unemployment from even more foreign competition, the Smoot-Hawley Tariff serves as an important reminder of the incredibly harmful aftershocks of economically nationalistic-driven policies. The Buy American Act is an example of a domestic content requirement or preference, an application of economic nationalistic attitudes and country-of-origin bias. As its name suggests, domestic content requirement and preference policies make it required or preferred for goods to be produced with goods sourced domestically. In similar vein to how domestic content requirements are utilized in Buy American (1933) 13 to protect American manufacturing and construction industries, India passed the Jawaharlal Nehru National Solar Mission (NSM), their own domestic content requirement aimed at strengthening their solar photovoltaic (PV) manufacturing base. Researchers Anshuman Sahoo and Gireesh Shrimali examine the Indian solar PV manufacturing sector’s competitiveness over time and cross referenced their findings with the sector’s trends in capacity utilization relative to others (2013). Their research found that India’s solar PV manufacturing sector was not only uncompetitive compared to related industries, but also experienced decreasing competitiveness relative to China’s competing industries. Additionally, they found that domestic developers in the solar PV sector were potentially favoring a different technology in order to bypass the domestic content requirement. Ultimately, the authors found that India’s solar PV sector struggled with innovations that their Chinese competitors did not. Domestic content requirements, thus, exacerbated the issue, sheltering the sector from potential synchronization with others. The impact of domestic content requirements on India’s solar photovoltaic manufacturing industry not only underscore the ineffectiveness of these measures, but also how they can insulate the issues underlying struggling industries. Following the global financial crisis of 2008, the United States and China passed stimulus packages that enacted domestic content requirements. Mario Larch and Wolfgang Lechthaler expand on the popularity of ‘Buy American’ or ‘Buy National’ rhetoric alongside their own analysis of the economic effectiveness of these measures. Through the documentation international protectionist activity by the International Monetary Fund (2009), World Bank (2009), and GlobalTradeAlert.org (2009), they 14 found that in comparison to the number of internationally implemented protectionist policies during the Great Depression, there were many fewer protectionist interventions following the 2008 global recession. Despite the significant decrease, however, The United States and China both enacted stimulus packages that incorporated ‘Buy National’ provisions, triggering international criticism and response (Larch & Lechthaler 2009). In order to examine the role of protectionism as a short-run response to recession, the authors focus on the dynamics of the economy while experiencing the recessional shock and the effects of protectionism during this state. Drawing from past work and their own model, Larch and Lechthaler find that protectionism hurts everyone – countries protecting themselves and those that the measures protect against. Even without trade retaliation, trade barriers foremost decrease the protecting nation’s imports. The protecting nation’s exports will also decline following a price level increase relative to its trading partners. This translates into reduced exporting firm business opportunities and higher consumer prices. Resources, thus, move from efficient exporting firms to their less productive counterparts (Larch & Lechthaler 2009). Despite these negative outcomes, Larch and Lechthaler find that governments still can default to protectionism for two primary reasons: first, non-exporting firms gain under the circumstances and second, closed economies can better shield from global economic shocks. In particular, the authors contend that the loss non-exporting firms feel relative to competitive exporting firms under a more liberated economy are felt more strongly. This notion is an example of loss aversion–the idea that losses are felt more than gains. Such attitude, when coupled with lobbying, can give politicians a 15 reason to strongly consider protectionism (Larch & Lechthaler 2009). Larch and Lechthaler’s argument is in line with previous works on economic nationalism: the perceived as threats and losses from foreign markets are a strong justification for protectionism. Despite their thorough criticisms of protectionism, Larch and Lechthaler bring up the positive role of protectionism in non-exporting firms. The multidimensionality of the economy inherently casts off black and white interpretations. Bai et al. expand on this notion using evidence from China’s industries (2002). They take a systematic approach to examine the role protectionism plays in economic efficiencies alongside scale economies–the idea that firms receive increasing returns with larger scale, and external economies–when these efficiencies are experiences in the industry rather than the single firm. Drawing from other sources, the authors used the average firm size in an industry served as a proxy for scale economies and the share of engineers and technicians in an industry’s employment was used for external economies. Using panel data from 32 two-digit industries from 29 Chinese regions over 13 years, Bai et al. found evidence supporting the use of local protectionism for industries with larger shares of state ownership. Compared with external scale and scale economies, protectionism played a larger role in bolstering efficiencies of these industries. On April 18, 2017, President Donald Trump signed an executive order reiterating the role of Buy American (1933) and similar laws with domestic content requirements. The order was an efforts of the President and his administration to fulfill their campaign promises of “reviving” the American manufacturing industry and jobs (Donald J. Trump for President, Inc 2019). In particular, the executive order called for 16 ensuring “the maximum utilization of goods, products, and materials produced in the United States” to the extent “permitted by law” (The White House, 2017). Economists’ reactions were immediate. Researchers with the Peterson Institute for International Economics (PIIE), Gary Clyde Hufbauer and Euijin Jung argue that two greatest and most significant losers from ‘Buy American’ are American taxpayers and the export industry, in line with the findings of Larch and Lechthaler (2009). Buy American (1933) and similar policies primarily involve federal procurement, how governments source inputs for public projects. Even before the negative effects are seen through empirical data, broad advocacy of Buy American policies risk the needed multilateral cooperation necessary for more open federal procurement provisions across trading partners. Drawing from the general notion that competition creates more consumer choice and environments for innovation, reducing the accessibility of open and competitive government procurement implies paying for lower quality and less innovative goods at a higher price. Even more so, Hufbauer and Jung argue Buy American has negative implications for U.S. exports. As noted by Larch and Lechthaler (2009) and Irwin (1996), trading partner retaliation not only compounds the negative effects of protectionism, but is generally expected. As major trading partners adopt similar policies, American exported goods and services face the same barriers in those governments’ procurement markets (Hufbauer & Jung 2017). Additionally, the authors note the many private business who not only largely make up construction exports, but also medical equipment, medicines, and information technology industries. They concede, however, that public procurement makes up a relatively small percentage of American goods and services exports. Even so, they 17 estimate a $189 billion export increase without the existence of Buy American and its international counterparts. On the contrary, I want to measure the change in imports as consequence of these policies as a proxy for American consuming habits. Like Hufbauer and Jung, trade economist, Tori Whiting argues that President Trump’s reinforcement of Buy American does significantly more harm than good. The policy is more pervasive than one might assume, affecting projects of the Federal Highway Administration, Federal Aviation Administration, Federal Transit Administration, Federal Railroad Administration, Amtrak, and through the Clean Water State Revolving Fund and the Drinking Water State Revolving Fund, Buy American affects the nation’s water supply as well (Whiting 2017). Her primary qualms with the domestic content requirements are threefold: they create additional regulatory hurdles for producers, bring additional cost to the American taxpayer, and are unlikely to yield targeted job growth in industries (2017). Contrary its purpose, Buy American has yielded consistent opposition from American businesses who cite the provisions as a detriment to their ability to compete in the market (Whiting 2017). Specifically, Whiting cites American steel manufacturing company, NLMK USA. The company uses steel slab to make rolled coil and galvanized steel products, employing about 1,100 Pennsylvania and Indiana residents. The company, however, must import most of its steel slab due to limited American supply; these products, in turn, are disqualified for U.S. highways, transit, and water projects (Whiting 2017). The Trade Partnership Worldwide estimates that 25 new $100,000 salary jobs could be created in Indiana alone if Buy American and similar domestic content rules were removed (Whiting 2017). In addition to stagnating job creation, Buy American’s preferential system of 18 adding from 6 to 50% to prices could cost taxpayers between an additional $53 million and $75 million (Whiting 2017). Finally, using employment and wage data from the Bureau of Labor Statistics, Whiting found that policies carrying Buy American provisions had no positive impact on the United States steel-manufacturing employment. She argues that removing domestic content requirements would increase U.S. GDP by $22 billion and create an estimated 363,000 jobs (while losing 57,000 jobs). Ultimately, Whiting argues that the Trump Administration and Congress ought focus on developing innovation-driven and job-creating economic environments over interventionist tactics. Whiting’s paper draws heavily from the work of Peter B. Dixon, Maureen T. Rimmer, and Robert G. Waschik of Victoria University’s Centre of Policy Studies. They use a tool called USAGE, a 389-industry computable general equilibrium model of the U.S. Economy developed at the Centre of Policy Studies, Victoria University with the U.S. International Trade Commission (Dixon et al. 2017). The authors simulate ‘Buy American’ policies through the assumption that U.S. industries artificially prefer to supply the U.S. government with domestically produced inputs of goods rather than those imported. By modeling the U.S. economy without American domestic content requirements, they model the policies’ negative repercussions. Using their economic model, they found a clear failure on the part of Buy American to promote aggregate employment and economic growth. The model indicated that industries targeted by domestic content requirements, especially iron and steel, were not strongly dependent on those policies. While they found that Buy American reduces manufacturing jobs by 0.439%, or 57,000 jobs, its implementation accounts for 9% of jobs (900 jobs) in light 19 fixture, plumbing materials, and wiring devices industries (2017). These findings are consistent with the general findings of Larch and Lechthaler (2009) and Bai et al. (2002), that certain industries are offered levels of protection under ‘Buy American’ legislation. Despite this, Dixon et al. contend that when weighing scrapping the Buy American Act and its counterparts, the winners heavily outweigh the losers; 50 out of 51 states (including Washington D.C.) and 430 out of 436 congressional districts would experience job growth. Moreover, abandoning these policies would hopefully be an international signal for the democratization of public procurement markets. While Dixon et al. utilize an incredibly thorough model (that took fifteen years to create), they acknowledge in their modeling of Buy American, that they did not find existing quantitative evidence on how input decisions by bidders to the U.S. government are biased against imports. Their observation underscores how majority of the quantitative research on the Buy American Act and similar domestic content requirements have not examined the role of ‘Buy American’ rhetoric on the choices of private firms. I will attempt to fill this gap. 20 Research Question This thesis examines the potential impact of “Made in America” rhetoric on American consumption of goods. Does it encourage American consumers to buy products sourced from the United States? How does this rhetoric have an affect U.S. importing behavior? Hypothesis I hypothesize that “Made in America” products will not translate into U.S. trade trends. Although many survey-based studies have indicated that country-of-origin and home biases play a statistically significant role in consumer preferences, these studies do not require participants to spend their actual dollars. Although these methods reveal legitimate preferences, they likely exaggerate the role of these biases. 21 Methodology Studies in the past have focused on surveying audiences to observe the role of country-of-origin and home bias effects. In his study of the impact of country-of-origin on product quality perceptions, Chao (1998) surveyed university students with images of products with a list of country-of-origin, price, and design specifications. Later research on country-of-origin effects emulate Chao’s survey methods, but test different consumer groups and accordingly adjust the specifications of the compared goods to align with their research scope. Costa et. al (2016) used an online survey to observe the perception of foreign goods in French consumers, comparing German and Brazilian made goods. Economists have also utilized bidding and auctioning survey methods in communities where it is commonly practiced when determining price differences consumers are willing to pay for different goods (Hoffman and Gatobu (2012); Morey (2016)). Surveying, however, does not require the participants to follow through with the transaction. As a result, results can overestimate significance of suggested biases. Looking at U.S. imports allows me to test if these biases are observable. Model To address my research question, I use a Difference-in-Differences (DID) model design, following similar steps to Morey (2016). The model is used to measure the effect of a “treatment”, a variable. DID has two different groups: a “treatment” group and “control” group. For my research, I focus on President Donald Trump’s executive order passed on April 18, 2017, titled “Buy American and Hire American.” Given this choice, the “treatment” group will be Buy American imports that are directly targeted 22 and affected while the “control” group is unaffected. I will detail the process of determining these two groups later in the methodology. First, I will explain why “Buy American and Hire American” is an attractive and effective “treatment” for the purposes of my research. Proxy of Economic Nationalism Economic nationalism strongly characterizes the context of the order’s signing. Passed in the first few months of Trump’s Administration, the executive order is emblematic of Donald Trump’s presidential campaign promises of bringing back stolen American jobs and reviving manufacturing industries. In the 2016 U.S. Presidential Campaign and well into his presidency, Trump has consistently painted threatening narratives concerning nations such as China, Germany, and Japan (Fisher 2018). His rhetoric surrounding manufacturing employment heavily parallels an economically nationalist zero-sum attitude Baughn and Yaprak (1996) describe, associating with economic gains and wins of trading partners as losses to the United States. President Trump’s choice to especially target the People’s Republic of China seamlessly mirrors Dana Frank’s (1996) analysis of the Yellow Perilism disguised in modern American economic nationalism. Thus, the executive order was one of the President’s first signs of putting those campaign promises in practice, stating that the order was a part of working “in our power to make sure more products are stamped with those wonderful words ‘Made in the USA,” because “for too long we’ve watched as our factories have been closed and our jobs have been sent to faraway lands” (Phillip 2017). 23 Price Effects Because “Buy American and Hire American” did not directly implement additional trading barriers or change the Buy American Acts existing provisions, its role is largely symbolic. Over year later in July 2018, the United States implemented tariffs valued at $34 billion on Chinese products, followed by a back-and-forth of tariff retaliations. These interventions have direct impact on import and export pricing, making discerning import biases much more difficult. Thus, having a “treatment” that causes minimal changes to existing prices helps me eliminate price effects and isolate consumer bias. Simplicity In January of 2019, President Trump signed another executive order titled “Strengthening Buy-American Preferences for Infrastructure Projects” as an extension of the first. Given its recency and the pace at which the United States Census Bureau publishes its import data, it made little sense to focus on it. Furthermore, the executive order passed amidst American and Chinese trade retaliations, which would play an even larger role in this hypothetical model. Furthermore, this executive order expanded Buy American to all factors of production, while “Buy American and Hire American” affected only steel and iron and articles thereof. A crucial part of the difference-in- differences model is the parallel trends assumption, where the control and treated groups should have parallel behaviors prior to the treatment. This means when looking for a goods that serve as the model’s “control” group, they must trend parallel to the import custom value of iron and steel, while not being covered by “Buy American and Hire American.” Goods that are most likely parallel to iron and steel are likely also 24 widely used in the realm of construction, making aluminum a favorable option. In economics, goods are substitutable if the price increase of one good causes demand increase for the other good. If the control and treated groups are substitutes, opposite movements following the treatment can indicate that “Buy American and Hire American” had an affect similar to a price level change. Figure 1 graphs U.S. monthly import levels for aluminum and steel. Figure 1: United States Iron & Steel and Aluminum Annual Imports Data Source: United States Census Bureau Visually, the two goods obey the parallel trends assumption. Thus, the two goods are likely not perfect substitutes (they do not have a perfect inverse relationship). If the treatment does have a significant effect on the two groups, the parallel trend between the treated and control group prior to treatment underscore a potential substitution effect. Because President Trump’s 2019 executive order also covers 25 aluminum, it eliminates that control group. Thus, we are given a unique window to analyze in 2017. Chosen Goods American imports are classified using the Harmonized Tariff Schedule of the United States Annotated (HTSA), a hierarchical structure for organizing goods for duty, quota, and statistical purposes (United States International Trade Commission 2019). The 2-digit HS code is the broadest form of product classification. Products are further specified by the 4 and 6-digit HS codes. The United States further specifies products by 8-digit rate lines and even more by 10-digit HS codes for statistical reporting purposes. Import data was collected from the United States Census Bureau’s USA Trade Online database which provides monthly customs values in U.S. Dollars up to the 10-digit HS code detail. GlobalTradeAlert.org is a website that tracks international interventions. If applicable, the website also records products effected at the 4-digit HS and 6-digit HS levels. Using its database, listed goods affected by Buy American interventions were chosen to be a part of the “treated” group at the 6-digit HS level. In total, I use 309 different commodities at the HS 6 level; iron and steel commodities make up 273 while aluminum commodities make up 36. This is due to the higher specifications given to iron and steel commodities. All commodities used and their respective 6-digit HS codes are listed in the Appendix. The U.S. Census Bureau provides data from as early as January 2001. Because aluminum becomes included in Buy American Act provisions after February 2019, my study ends after that date. To keep the periods before and after the treatment symmetric, the data set includes observations of 20 monthly periods 26 before and after the treatment. Thus, the observations include data from June of 2013 to January of 2019. Table 1 gives summary statistics of the custom values and logged custom values. Note that values for iron and steel are the sum of the total import values for 2-digit HS codes 72, titled “Iron and Steel”, and 73, titled “Articles Of Iron Or Steel”, while the value for aluminum is the total import values for 2-digit HS code 76, titled “Aluminum And Articles Thereof.” Summary statistics Entire sample Iron & Steel Aluminum Total Custom Value 460,527 354,069 106,457 (in hundred thousand of US$) Average Monthly Customs Value 6,978 5,365 1,613 (in hundred thousand of US$) Observations 20,164 17,854 2,310 Table 1: Observation Summary Statistics Model Specifications Given the variables selected, my most rudimentary model is as follows: LoggedCustomValueHS, t = β0 + β1TreatmentHS, t + β2TreatedHS, t + β3 (TreatmentHS, t)(TreatedHS, t) + ε HS,t Where subscript HS refers to the HS code at the 6-digit level and the subscript t refers to time. The variable Treatment is a dummy variable for President Trump’s executive order that turns on and after April 2017. The variable Treated is a dummy variable that turns on for HS 6-digit level products that GlobalTradeAlert.org marks as affected. The 27 interaction variable, Treatment*Treated measures the difference in differences. Its estimated β3 constant models the impact of Buy American as follows: ImpactBuy American = Δ(ImportHS,t - ImportHS, t-1) Where t indicates the treatment and t-1 indicates the period(s) before treatment and subscript HS still refers to the HS code at the 6-digit level. Software All statistical analysis was conducted within STATA. Data was downloaded from the U.S. Census Bureau into csv files where I reformatted for STATA’s reading purposes. My entire STATA code is listed in the Appendix. 28 Results I ran a preliminary level regression to give me an early idea of what the data would predict. Results of an unlogged difference in differences regression are located in the Appendix. The coefficient on TreatmentxTreatedGroup is the primary indicator if “Buy American and Hire American” influences U.S. imports. Iron and steel and their articles thereof, which are affected by the Buy American Act, have a negative relationship with levels of custom values imported (p=0.000). In general, all goods were positively affected by the treatment—President Trump’s executive order and public announcements of strengthening of Buy American provisions. All included goods were imported at higher levels after Trump’s announcement (p=0.000). The interaction variable between commodities affected by Buy American and President Trump’s executive order has a negative coefficient, suggesting that Buy American affected products were imported at lower levels than unaffected products after the signing of the executive order (p=0.000). The coefficient values are so large, however, making interpretation very difficult. The extremely high root mean square error alongside the very significant p- values (p=0.000) are just a few signs that level data is not the best choice. For the rest of my regressions, I take the log of the custom values. Not only does this correct the skewness from the import custom values being in the billions, it allows for percentage change interpretations. The timet variable is added to eliminate time invariant fixed effects. Additionally, I add control variables to account for events affecting certain commodities or at certain times that otherwise would be absorbed in the error term. The variable USChinaTradeWar is a dummy variable that turns on after March 2018, when 29 the United States imposed a 25 percent tariff on all steel imports (excluding Argentina, Australia, Brazil, and South Korea) and a 10 percent tariff on all aluminum imports (excluding Argentina and Australia) (Wong & Koty 2019). The variable DutyInvestigationsHS, t is a dummy variable that turns on for commodities under the 4- digit HS code 7208 and after June 2018, when the U.S. began countervailing and antidumping duty investigations of steel racks imported from China (GlobalTradeAlert.org). The variable CanadaMoniteringHS, t is a dummy variable that turns on for all commodities under the 2-digit HS code 72 and 73 and after November 2017, when Global Affairs Canada announced to Importers the extension of their import monitoring program (GlobalTradeAlert.org). The variable ChineseTariffsHS,t is a dummy variable that turns on for commodities under the 4-digit HS codes 7204 and 7602 and after August 2018, indicating when the Chinese government announced $16 billion worth of tariffs. These controls and fixed effects are added into the regressions seen on Table 2. Regression (1) is the simplest regression, emulating the preliminary regression ran but with logged custom values. Table 2: Difference in differences regression results (not robust) Logged Custom ValueHS, t (1) (2) (3) (4) (TreatmentHS, t)* -0.077 * -0.077 * -0.077 * -0.067 (TreatedGroupHS, t) (0.041) (0.041) (0.041) (0.120) TreatmentHS, t 0.0936 ** 0.166 *** 0.162 *** 0.158 *** (0.008) (0.000) (0.000) (0.000) TreatedGroupHS, t -0.297 *** -0.297 *** -0.297 *** -0.297 *** (0.000) (0.000) (0.000) (0.000) 30 - - - timet 0.0000072 *** 0.000074 *** 0.000074 *** (0.894) (0.352) (0.202) USChinaTradeWart 0.0108 0.0140 (0.601) (0.588) DutyInvestigationsHS, t -0.00915 (0.905) CanadaMonitoringHS, t -0.0130 (0.663) ChineseTariffsHS, t 0.282 * (0.016) constant 7.002 *** 8.452 *** 8.499 *** 8.496 *** (0.000) (0.000) (0.000) (0.000) N 20164 20164 20164 20164 R-squared 0.017 0.018 0.018 0.018 p-values in parentheses * p<0.05, ** p<0.01, ***p<0.001 Regression (2) includes a variable to account for time fixed effects. Regression (3) includes the dummy variable for the start of the U.S.-Chinese Trade War. Regression (4) adds in the rest of the control variables. As mentioned before, the interaction variable, (TreatmentHS, t)*(TreatedGroupHS, t) is the key to measuring the effect of the Buy American Act on goods affected. We can see that even as fixed effects and explanatory variables are being added, the coefficient on the interaction variable remains relatively the same while the p-value decreases. 31 Heteroskedasticity and Multicollinearity To ensure that my error terms are constant, I conducted the White Test for heteroskedasticity on Regression (4). The results are shown in Figure 2. Figure 2: White’s Test for heteroskedasticity With a p-value = 0.2916, the null hypothesis of homoskedasticity is not rejected, thus the OLS assumption of homoskedasticity is retained. Additionally, because the model has multiple dummy variables, I want to look at multicollinearity to ensure that the model does not account for effects multiple times. Table 3: Multicollinearity between variables Looking at Table 3, notice that the CanadaMoniteringHS, t has consistently higher correlations with other variables. With this in mind and to further ensure that the standard error values are accurate, I reran the regressions with robust standard errors and including a regression without the variable, CanadaMoniteringHS,t. The results are in Table 4, which displays p-values in parentheses, and Table 5, which alternatively displays standard errors. 32 Table 4: Robust OLS regression results with p-values 33 Table 5: Robust OLS regression results with standard errors We can see that standard errors uniformly increase with the addition of CanadaMoniteringHS, t, so it will be excluded for the final regression. The formula for the final regression is as follows: LoggedCustomValueHS, t = β0 + β1TreatmentHS, t + β2TreatedHS, t + β3 (TreatmentHS, t)(TreatedHS, t) + β4USChinaTradeWarHS, t + β5 DutyInvestigationsHS, t + β6 ChineseTariffsHS, t + ε HS, t Its regression results are in Table 6. 34 Table 6: Final OLS regression results Robustness Checks Another potential factor that could be skewing results are the chosen cutoff dates. Tables 7 and 8 runs the final model with different time periods; Table 7 shows p- values in parentheses while Table 8 shows standard errors in parentheses. Regression (1) is the original regression with 20 periods before and after the treatment. Regression (2) extends the period prior to treatment by an extra year, or 12 periods; regression (3) extends the period prior to treatment by only 3 periods; regression (4) reduces the period prior to treatment by 3 periods; regression (5) reduces the period prior to treatment by an extra year, or 12 periods. Across the board, p-values and standard errors for the (TreatmentHS, t)*(TreatedGroupHS, t) coefficient seem to be inversely related to the number of periods before the passage of President Trump’s executive order. 35 Table 7: Robustness checks, p-values The increasing p-values and standard errors are likely the result of the Difference-in- Differences model parallel trend assumption breaking down as fewer data points are used for the regression. Looking at the Figure 1, the difference between level custom imports is less uniform. This can detract from the validity of the model and must be critically considered when interpreting regression results. 36 Table 8: Robustness checks, standard errors Discussion When comparing the results from Table 7 to Table 1, notice that all coefficients maintained their positive or negative states, and thus all relationship interpretations remain the same. The p-value on the first three variables are all statistically significant below the five percent level. Robust standard errors and the root mean squared error (Root MSE) also sit at low values. Most significantly, the results indicate that Buy American affected goods experienced a 7.55 percent decrease (p=0.047) after President Trump signed the executive order, “Buy American and Hire American.” Import levels 37 of all goods, however, increase by 16.4 percent (p=0.000) following President Trump’s executive order, while iron and steel imports in general, are imported 29.9 percent less (p=0.000) relative to the imports of aluminum. Considering that iron and steel made up most the observations, it is surprising that despite the coefficient for the treated group being negative, that the treatment’s overall correlation with logged custom value is positive. While interpreting these results, it is important to note that the Buy American Act and President Trump’s executive order do not raise the prices of foreign products in all markets. Rather, only the United States government faces the artificially increased prices. Furthermore, Trump’s executive order did not change the Buy American policies of the status quo, but merely recalled the importance of such measures. But unlike the private construction market, government spending is not only restrained by a tax revenue, but a slow growing tax revenue. Thus, when the government is artificially increasing its prices, its ability to invest in construction and infrastructure projects also significantly decreases due to the budget constraint. Accordingly, there are two simple narratives that explain the negative relationship between the Buy American Act enforcement and the goods it covers. First, consumers responded positively to President Trump’s support of buying domestic and emulated it in their own buying behavior. This explanation seems unlikely due to the overall statistically significant positive relationship between the executive order and imports. Furthermore, American steel company, NLMK USA, serves as anecdotal evidence that steel imports are driven largely by necessity as a reality of limited American supply, rather than a hardline for 38 supporting domestically sourced goods. Thus, a likelier explanation is the government’s reduction in its own ability to import commodities due to budgetary constraints. Then what accounts for the positive increase in imports overall after a politically and economically significant example of economic nationalism? Potentially those who are not constricted by the taxpayer’s dollar, private construction, raised overall imports. Figure 3 shows the monthly levels of private versus public construction spending over the last decade. Figure 3: United States Public and Private Construction Spending Data Source: United States Census Bureau Since the 2008 recession, private construction spending has both increased higher and at faster rates than public construction spending. As major factors of construction, it logically follows that import levels increase for iron, steel, and aluminum. 39 Figure 4: United States Infrastructure Spending Source: Peter G. Peterson Foundation This fact, coupled with the reality of decreasing government infrastructure spending, depicted in Figure 4, can potentially explain demand increases for construction related commodities and imports. Model Limitations Although I previously cited simplicity as a reason behind choosing “Buy American and Hire American” and the Buy American Act as the treatment for this model, the original act and its executive order counterpart come with many limitations. First, because the policies exclusively affect public procurement, it precludes the 40 research from observing potential economically nationalist biased non-governmental consumption. Furthermore, although my dataset contains many substitutable good sets at the 4-digit and 6-digit HS levels, they are all articles or types of iron and steel or aluminum. In particular, these metals face a sea of barriers and are often the commodity of choice when weaponizing trade relationships. As a result, the turbulent political environment can make modeling around these commodities difficult. The difference in differences model, however, is very helpful in this regard, as for the most part, aluminum, iron, and steel products are often targeted together. Another caveat in the data is the sole use of custom values imported to the United States. Having access to the units imported would have enabled calculations for imports’ average weight and measure the impact of the treatment on imported units. Additionally, it could provide better insight to the relationship between the control and treated groups. This information was limited by the United States Census Bureau database. Such additional measures and regressions could have provided better narrative insight and context and should be explored in further research of this topic. 41 Conclusion Although economic nationalism, consumer biases, and domestic content requirements have been extensively researched in a variety of fields, the majority of research focuses on either the individual preferences of consumers without budget constraints or on the specific policy interventions that have direct price implications. Since economic nationalism and home bias has been shown to be a significant influence in consumer behavior, this paper has attempted to model the effect of a representative case of economic nationalism on American consumption of foreign goods–imports. Using an OLS regression to estimate a difference in differences, model results show that the representative case, President Trump’s first executive order, “Buy American and Hire American,” correlated with an approximately 10 percent reduction in imports for goods covered by Buy American Act provisions, but also about an 8 percent increase in imports overall. Looking at these results alongside the stagnant levels of U.S. government construction spending and the rising private construction spending, suggests a narrative less patriotic than the slogan of the executive order would imply. As the government artificially increases the prices of construction inputs to accommodate for domestic prices while working within the budget of the taxpayer dollar, the budget constraint suggests a lower ability to invest in public infrastructure. This fact is reinforced by declining U.S. infrastructure spending. Alongside previous Buy American Act research by Hufbauer and Jung (2017), Tori Whiting (2017), and Dixon et al. (2017), this paper finds that these domestic content requirements also work against price-driven consumers. It is the U.S. economy, not the government, that is responsible for job creation, while the government’s role is 42 to “create a tax, trade, and regulatory environment where private businesses are able to grow and flourish” (Whiting 2017). Domestic content requirements such as the Buy American Act and “Buy American and Hire American” are harmful economically nationalist policies that ultimately pander to party allegiances while harming the American public. 43 Appendix A: STATA Code import delimited "/Users/melissaliu/iCloud Drive (Archive)/Desktop/Thesis Work/alum vs steel .csv", clear // Destringing US Census's time formatting and making it Stata comaptible gen time1 = date(time, "MDY") /* Dummy variable that turns on after April 2017, when President Trump signs executive order, "Buy American and Hire American */ generate Treatment = (time1 > 20545) // The following are dummies for fixed effects and control variables generate USChinaTradeWar = (time1 > 21216) generate DutyInvestigations = time1 >= 21336 & hs4==7208 generate CanadaMonitoring = time1 >= 21124 & hs2 <= 73 generate ChineseTariffs = time1 >=21397 & hs4==7204 | time1 >=21397 & hs4==7602 // Generating interaction variable between 'treatment' and 'treated group' gen TreatmentxTreatedGroup = Treatment*treatedgroup // Establishing the time periods drop if time1 > 21550 drop if time1 < 19571 // Summary Table of Iron and Steel Imports summarize customsvaluegenus logcustomsvalue if treatedgroup == 1 44 // Summary Table of Aluminum Imports summarize customsvaluegenus logcustomsvalue if treatedgroup == 0 ssc install estout, replace eststo clear eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar DutyInvestigations CanadaMonitoring ChineseTariffs esttab, p r2 // Breusch-Pagan tests of hetereoskedasticity hettest TreatmentxTreatedGroup Treatment treatedgroup hettest TreatmentxTreatedGroup Treatment treatedgroup time1 hettest TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar hettest TreatmentxTreatedGroup Treatment treatedgroup time1 DutyInvestigations CanadaMonitoring ChineseTariffs hettest TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar DutyInvestigations CanadaMonitoring ChineseTariffs 45 eststo clear eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup, robust eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1, robust eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar, robust eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar DutyInvestigations ChineseTariffs, robust eststo: quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar DutyInvestigations ChineseTariffs CanadaMonitoring, robust esttab, p r2 corr logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar DutyInvestigations CanadaMonitoring ChineseTariffs, means quietly reg logcustomsvalue TreatmentxTreatedGroup Treatment treatedgroup time1 USChinaTradeWar DutyInvestigations CanadaMonitoring ChineseTariffs estat imtest, white 46 Appendix B: Preliminary Regression Results The following table are the regression results from running the model: CustomValueHS, t = β0 + β1TreatmentHS, t + β2TreatedHS, t + β3 (TreatmentHS, t)(TreatedHS, t) + ε HS,t 47 Appendix C: Commodities 6-Digit HS Code Product Description Effected by BAA 720110 Nonalloy Pig Iron 0.5 Prcnt Or Less Phosphorus yes 720150 Alloy Pig Iron; Spiegeleisen, In Primary Forms yes 720211 Ferromanganese With Over 2 Percent Carbon yes 720219 Ferromanganese, 2 Percent Or Less Carbon yes 720221 Ferrosilicon With Over 55 Percent Silicon yes 720229 Ferrosilicon, 55 Percent Or Less Silicon yes 720230 Ferrosilicon Manganese yes 720241 Ferrochromium Over 4 Percent Carbon yes 720249 Ferrochromium, 4 Percent Or Less Carbon yes 720250 Ferrosilicon Chromium yes 720260 Ferronickel yes 720270 Ferromolybdenum yes 720280 Ferrotungsten And Ferrosilicon Tungsten yes 720291 Ferrotitanium And Ferrosilicon Titanium yes 720292 Ferrovanadium yes 720293 Ferroniobium yes 720299 Ferroalloys, Nesoi yes 720410 Cast Iron Waste And Scrap yes 720421 Stainless Steel Waste And Scrap yes 720429 Alloy Steel Waste And Scrap, Not Stainless yes 720430 Tinned Iron Or Steel Waste And Scrap yes 720441 Ferrous Waste & Scrap Nesoi, Turnings, Chips Etc yes 720449 Ferrous Waste & Scrap Nesoi yes 720711 Smfd Ios Na U.25PCT Crbn Rec/sq Cs Wdth Un 2x Thns yes 720712 Smfd Irn/nal Stl Lt .25 Pct Crb Rect Cs Wid 2x Thk yes 720719 Smfd Irn/nal Stl Lt 0.25 Pct Carbon Cs Nt Rect/sqr yes 720720 Smfd Irn Or Nonalloy Stl, .25 Pct Or More Carbon yes 720810 Flat-hot-roll Iron, Nonaly Stl,coils,pttrns, Nes0i yes 720825 Flat-hot-roll Irn,nonaly,coil,pkled,4.75mm >,nesoi yes 720826 Flt-hot-rol Irn Nonaly,coil,pkld,3mm <4.75mm,nesoi yes 720827 Flt-hot-rol Irn,noaly Stl,coil,pk,<3mm Thick,nesoi yes 720836 Flt-hot-roll Irn,nonaly Stl,coil,>10mm Thick,nesoi yes 720837 Flt-hot-rol Irn,nonaly,coils,4.75mm N/o 10mm Nesoi yes 720838 Flt-ht-rl Irn,noaly Stl,coil,3mm But < 4.75MMNESOI yes 720839 Flat-hot-roll Irn,nonaly Stl,coil,<3mm Thick,nesoi yes 720840 Fr Ios Nal 600mm Ao Hr Nt C/p/c/cls Pttrns In Rel yes 720851 Flt-hot-roll Irn,nonaly St,not Coil>10mm Thk,nesoi yes 720852 Fr Ios Nal 600mm Ao Hr Nt C/p/c/cls 4.75-10MM Thck yes 720853 Fr Ios Nal 600 Ao Hr Nt C/p/c/cls 3-un 4.75MM Thck yes 720854 Fr Irn/nal Stl 600mm Ao Hr Nt C/p/c/cls Un 3mm Thk yes 720890 Fr Ios Na Cornc 600mm Ao W Hr Pl Nesoi yes 720915 Flt-cold-rol Irn,noaly,coil,600mm Wide,3mm > Thick yes 720916 Fl-cld-rl Irn,st,coil,600mm Wide,>1mm But <3mm Thk yes 720917 Fl-cld-rl Irn,st,coil,600mm Wd,0.5mmbut N/o 1mm Tk yes 48 6-Digit HS Code Product Description Effected by BAA 720918 Flat-cold-rld Ir,stl,coils,600mm Wide,<0.5mm Thick yes 720925 Flt-cld-rld Ir,st,not Coil,600mm Wide,3mm Or > Thk yes 720926 Flt-cld-rld Ir,st,not Coil,600mm Wd, >1mm <3mm Thk yes 720927 Flt-cld-rld,not Coil 600mm W,>0.5mmbut N/o 1mm Thk yes 720928 Flt-cld-rld Ir,nonal,notcoil,600mm Wide,>0.5mm Thk yes 720990 Fr Ios Na Cls Or Nt 600mm Ao W Cr Pl Nesoi yes 721011 Fr Ios Na 600mm Ao W Tin Coatd Or Pltd 0.5MM Ao Th yes 721012 Fr Ios Nal 600mm Ao W Tin Ctd Or Pltd Undr 0.5MM T yes 721020 Fr Ios Na 600mm Ao W Lead Ctd Or Pltd (ternplate) yes 721030 Flat-rld Iron,nonal Stl,600mm Wide,elec Platd Zinc yes 721041 Fr Ios Na 600mm Ao W Zinc Ctd O Pltd Nt Elctc Corr yes 721049 Fr Ir/nas Ctd/pltd W Zinc Nt Elec Nt Corr 600mm Om yes 721050 Fr Ios Na 600mm Ao W Ctd/pltd W Cro Or Cr And Cro yes 721061 Fr Iron/nonalloy Steel 600mm Ao,pltd/ctd Alum-znc yes 721069 Fr Iron/nonalloy Steel,600mm Ao,pltd/ctd Othr Alum yes 721070 Fr Ir/nas 600mm W Om, Painted, Varnished, Plastic yes 721090 Flat-roll Iron Or Nonalloy Steel Nu600mmclad Nesoi yes 721113 Fr Hs Ios Na Un600mm W Hr Pl Unvrsl Mllplte yes 721114 Fr Hs Ios Na Un600mm W Hr Pl 4.75MM Ao Thck yes 721119 Oth Fr Hi-str St Un 600mm W Npld Un4.75mm Thck yes 721123 Flat-cold-rolled Iron,nonalloy Steel, <600 mm Wide yes 721129 Flat-rolled Iron/nonalloy Steel Undr 600 mm, Nesoi yes 721190 Fr Ios Na Un 600mm W, Nesoi Mr Thn C-r yes 721210 Fr Ios Na Undr 600mm Wide Tin Coatd O Plated yes 721220 Flat-rld Irn,nonal St, < 600mm Wide Elec Pltd Zinc yes 721230 Flat-rld Ir/nas Un 600mm W Pltd/ctd W Zinc Nt Elec yes 721240 Flat-rld Ir/nas Un 600mm W Pntd/varnshd/plstc Ctd yes 721250 Fr Ios Na Undr 600mm Wide Pltd Or Coatd, Nesoi yes 721260 Fr Ios Un 600mm Wd Clad Nesoi yes 721310 Bars And Rods Irregular Coils Concrete Reinforcing yes 721320 Brs Rods Hot-rlld Irreg Coils Free-cuttng Steel yes 721391 Bars,rodshot-roll,irnnoal St Coil Circ,<14mm Nesoi yes 721399 Bars,rods,hot-rolled,iron Or Nonal Stl,coils,nesoi yes 721410 Other Bars And Rods Iron Or Nonalloy Steel, Forged yes 721420 Oth Brs Rds Ios Na Hot-wrkd, Conc Reinfrcng yes 721430 Other Bars And Rods Free-cutting Steel, Hot-worked yes 721491 Bars,rods,hot-rolled,-drawn,-ext,rectangular,nesoi yes 721499 Bars,rods,irn,noal,hot-rolled,-drawn-extrude,nesoi yes 721510 Oth Brs And Rds Free-cttng Stl Cold-fmd Or Fnshd yes 721550 Bars,rods,irn,noal,cold-formed,cold-finished,nesoi yes 721590 Bars And Rods Iron Or Nonalloy Steel, Nesoi yes 721610 U-i-h-sections Ir/nas Hot/wrkd Ls Thn 80mm High yes 721621 L Sec Ios Na Hot-wkd Lss Th 80mm High yes 721622 T Sec Ios Na Hot-wkd Lss Th 80mm High yes 721631 U Sec Ios Na Hot-wkd 80mm Or More High yes 721632 I Sec Ios Na Hot-wkd 80mm Ao High (standard Beams) yes 721633 H Sections Irn/nas, Hot-wrkd, 80mm Hi Or More yes 49 6-Digit HS Code Product Description Effected by BAA 721640 L Or T Sections Ir/nas Hot-wrkd, 80mm Hi Or More yes 721650 Oth Angls Shps Sec Ios Na Hot-wkd yes 721661 Angls Shps Sec Ir/nas Nt Frthr Cld-wrkd Frm Fr Pro yes 721669 Angls Shps Sec Ir/nas Nt Frthr Cld-wrkd Nt Fr Prod yes 721691 Angls Shps Sec Irn/nas Oth Cld-wrkd Fr Fr Products yes 721699 Angles Shapes Sections Iron/nonalloy Steel Nesoi yes 721710 Wire Of Iron Or Nonalloy Stl,not Plated Or Coated yes 721720 Wire Of Iron,nonaly Stl,plated Or Coated With Zinc yes 721730 Wire Of Irn,noal St, Plted Or Ctd Base Metal,nesoi yes 721790 Wire Iron/nonalloy Steel Under 0.25% Carbon, Nesoi yes 721810 Stainless Steel Ingots And Other Primary Forms yes 721891 Smfnshed Prdcts,stainless Steel,rectngulr (nt Sqr) yes 721899 Other Semifinished Products Of Stainless Steel yes 721911 Fr ss 600mm Ao W Hr Cls Ov 10mm Thck yes 721912 Fr ss 600mm Ao W Hr Cls 4.75-NOV 10mm Thck yes 721913 Fr ss 600mm Ao W Hr Cls 3-un 4.75MM Thck yes 721914 Fr ss 600mm Ao W Hr Cls Un 3mm Thck yes 721921 Fr ss 600mm Ao W Hr Nt Cld Ov 10mm Thck yes 721922 Fr ss 600mm Ao W Hr Nt Cld 4.75-NOV 10mm Thck yes 721923 Fr ss 600mm Ao W Hr Nt Cld 3-un 4.75MM Thck yes 721924 Fr ss 600mm Ao W Hr Nt Cld Undr 3mm Thck yes 721931 Flt-rld Stnls Stl 600mm Om W Cld/rld 4.75MM Om Thk yes 721932 Flt-rld Stnls Stl 600mm Om W Cld-rld 3-un4.75mm Th yes 721933 Fl-rld Stnls Stl 600mm Om W C-r Ov 1mm Un3mm Thk yes 721934 Flt-rld Stnls Stl 600mm Om W Cld-rld .5-1 mm Thck yes 721935 Flt-rld Stnls Stl 600mm Om W Cld-rld Un.5 mm Thick yes 721990 Flt-rld Stnls Stl 600mm Ao Wide, Nesoi yes 722011 Fr ss Undr 600mm W Hr 4.75MM Ao Thck yes 722012 Flt-rld Stnls Stl Un 600mm Wide Ht-rld Un4.75mm Th yes 722020 Flat-rolled Stnls Stl Und 600mm Wide, Cld-rld yes 722090 Fl-rld Stnls Stl Un 600mm Wde, Nesoi yes 722100 Bars And Rods, Stnls Stl, Ht-rld, Irreg Coils yes 722211 Oth Bars & Rods Stainless Steel Circ Cross-section yes 722219 Oth Bars & Rods Stnless Steel Hr Nt Circ Cross-sec yes 722220 Oth Bars A Rods, Stnls Stl, Nt Fur Th Cld-frm/fnsh yes 722230 Other Bars And Rods Stainless Steel, Nesoi yes 722240 Angles, Shapes And Sections Of Stainless Steel yes 722300 Wire Of Stainless Steel yes 722511 Flt-rlld Grain-ornted Silicon Elctrcl Stl 600mm Ao yes 722519 Flt-rld Silicon Elctrcl Stl 600mm Ao Nt Grain-ornt yes 722520 Flat-rolled High-speed Steel 600mm Wide Or More yes 722530 Flt-rld Oth Alloy Stl 600mm Om W, Ht-rld, Coils yes 722540 Flt-rld Oth Alloy Stl 600mm Om W, Ht-rld Nt Coils yes 722550 Flt-rld Oth Alloy Stl 600mm Om W, Cld-rld yes 722591 Flt-rld Alloy Stl Nesoi 600ao Elctlyc Plt/ctd Zinc yes 722592 Flt-rld Alloy Stl Nesoi 600ao Plt/ctd Zinc Nt Elct yes 722599 Flt-rld Alloy Steel Nt Stainless 600mm Ao W Nesoi yes 50 6-Digit HS Code Product Description Effected by BAA 722611 Flt-rlld Silicon Elctrcl Steel Un 600mm W Grain-or yes 722619 Flt-rld Silicon Elctrcl Stl Un 600mm W Nt Grain-or yes 722620 Flat-rolled High-speed Steel Under 600mm Wide yes 722691 Flat-rolled Oth Alloy Stl Un 600mm W Ht-rld yes 722692 Flt-rld Oth Alloy Stl Un 600mm W, Cld-rld yes 722694 Fr Alloy Stl Nes Un 600mm Pltd/ctd Zinc Nt Elctlyt yes 722699 Flat-rolled Other Alloy Steel Un 600mm W Nesoi yes 722710 Brs A Rds Hspd Stl Irrg Coils Hot-rolld yes 722720 Brs A Rds Slco-mn Stl Irrg Coils Hot-rolld yes 722790 Bars And Rods Oth Alloy Stl, Hot-rld, Irreg Coils yes 722810 Other Bars And Rods Of High-speed Steel yes 722820 Other Bars And Rods Of Silico-manganese Steel yes 722830 Oth Brs A Rods Oth Aly Stl Nt Fur Th Ht-rld/drn/ex yes 722840 Oth Brs Rds Oth Alloy Stl Nt Frthr Wrkd Thn Forged yes 722850 Oth Brs A Rods Oth Aly Stl Nt Fur Th Cld-frmd/fnsh yes 722860 Other Bars And Rods Of Other Alloy Steel, Nesoi yes 722870 Angles, Shapes And Sections Of Other Alloy Steel yes 722880 Hollow Drill Bars And Rods, Ios, Nesoi yes 722910 Wire Of High-speed Steel yes 722920 Wire Of Silico-manganese Steel yes 722990 Wire Of Other Alloy Steel, Nesoi yes 730110 Sheet Piling Of Iron Or Steel yes 730120 Welded Angles, Shapes And Sections, Iron Or Steel yes 730210 Railway Or Tramway Rails Of Iron Or Steel yes 730230 Swtchblds, X-ing Frgs, Pt Rds And Oth Ios Xing Pcs yes 730240 Fish-plates And Sole Plates Of Iron Or Steel yes 730290 Railway Or Tramway Track Constr Matrl Of Ios Nesoi yes 730300 Tubes, Pipes And Hollow Profiles Of Cast Iron yes 730410 Line Pipe For Oil Or Gas Lines Nsm, Ir Nesoi Steel yes 730421 Drill Pipe Used For Oil,gas Drilling,iron Or Steel yes 730429 Casing And Tubing,oil,gas Drilling, Iron Or Steel yes 730431 Oth Ios Na Ps Tb Hlw Pfl Smls Cir Cs Cold-wrkd yes 730439 Oth Ios Na Ps Tb Hlw Pfl Smls Cir Cs Nt Cld-wrkd yes 730441 Oth ss Tb Ps Hlw Pfl Smls Circ Cs Cold-wrkd yes 730449 Tubes, Pipes Etc Nesoi, Circ Cr Sect, Stainless St yes 730451 Oth Tb Ps Hlw Pfl Aly Stl Nt ss Smls Circ Cs Cd-wk yes 730459 Oth Tb Ps Hp Aly Stl Nt ss Smls Circ Cs Nt Cld-wrk yes 730490 Tubes, Pipes Etc, Seamless Nesoi, Ir Nesoi & Steel yes 730511 Pipe, Oil Line Etc Ov16in Ir St, Long Sub Arc Weld yes 730512 Pipe, Oil Line Etc Ov16in Ir Or St, Longi Wd Nesoi yes 730519 Pipe, Oil Line Etc Ov16in Ir Or Steel, Close Nesoi yes 730520 Casing, Oil Or Gas Drillng Over16in, Iron Or Steel yes 730531 Pipe Nesoi, Ov16in Iron Or Steel, Longit Welded yes 730539 Pipe Nesoi, Ov16in Iron Or Steel, Welded Nesoi yes 730590 Pipe Nesoi, Ov16in Iron Or Steel, Closed Nesoi yes 730610 Pipe For Oil Or Gas Pipelines Iron Or Steel Nesoi yes 730620 Casing Etc Oil Or Gas Drillng, Iron Or Steel Nesoi yes 51 6-Digit HS Code Product Description Effected by BAA 730630 Pipe Etc Nesoi, Weld Cir Cr Sect, Iron Or Nonal St yes 730640 Pipe Etc Nesoi, Weld Cir Cr Sect, Stainless Steel yes 730650 Pipe Etc Nesoi, Weld Cir Cr Sec, Alloy Steel Nesoi yes 730660 Pipe Etc Nesoi, Weld Noncir Cr Sec, Iron Or Steel yes 730690 Pipes Etc Nesoi, Riveted Etc, Of Iron Or Steel yes 730711 Cast Pipe Fittings, Nonmalleable Cast Iron yes 730719 Cast Pipe Fittings, Iron Nesoi Or Steel yes 730721 Pipe Or Tube Fittings Nesoi, St Steel Flanges yes 730722 Pipe Fittings Nesoi, Stainless Steel Thr Elbow Etc yes 730723 Stainless Steel Tube Or Pipe Butt Welding Fittings yes 730729 Stainless Steel Tube Or Pipe Fittings Nesoi yes 730791 Pipe Fittings Nesoi, Iron Or Nonst Steel Flanges yes 730792 Pipe Fittings Nesoi, Ir Or Nonst St Thr Elbows Etc yes 730793 Pipe Fittings Nesoi, Ir Or Nonst St, Butt Weld Fit yes 730799 Pipe Fittings Nesoi, Of Iron Or Nonst Steel Nesoi yes 730810 Bridges And Bridge Sections Of Iron Or Steel yes 730820 Towers And Lattice Masts Of Iron Or Steel yes 730830 Drs, Wndws A Frms A Thrshlds Fr Drs, Iron Or Steel yes 730840 Equip For Scafldg/shuttrg Proppg/pit-proppg Ir/stl yes 730890 Structures And Parts Nesoi Of Iron Or Steel yes 730900 Tanks Etc, Over 300 Liter Capacity, Iron Or Steel yes 731010 Tanks Etc, Capacity Notun50notov300 Liter, Ir & St yes 731021 Cans To Be Soldered/crimped Closed Ir/st Un 50 Ltr yes 731029 Tanks Csks Drms Cns Bxs Etc Ios Nesoi Und 50 Ltr yes 731100 Contnrs Fr Cmprssd O Lqfd Gas Of Iron O Steel yes 731210 Stranded Wire, Rope Etc, No Elect Insul, Ir Or St yes 731290 Plaited Bands, Slings Etc, Iron Or Steel Nesoi yes 731300 Barbed Wire And Twisted Wire For Fencing, Iron/stl yes 731412 Endless Bands Of Stainless Steel yes 731413 Endlss Bnds,wovn Iron/steel Wire Clth,nt Stainless yes 731414 Other Products Of Woven Stainless Steel Cloth yes 731419 Woven Products Iron Or Steel, Nesoi yes 731420 Grill Netting Fencing Wld Ir/st Wr 3mmcs 100cm2msh yes 731431 Oth Grll Nttng A Fncng Wldd At Intrsct Galvnzed St yes 731439 Oth Grll Nttng A Fncng Wldd At Intrsct Ios Nt Glvn yes 731441 Oth Grill, Nettg Fncg Ios Ctd/pl W Zn Nesoi Nt Wld yes 731442 Grill Netting Fencing, Plastic Coated Ios Wr Nesoi yes 731449 Oth Grill, Nttng Or Fncng Nesoi Of Iron Or Steel yes 731450 Expanded Metal, Iron Or Steel yes 731511 Roller Chain Of Iron Or Steel yes 731512 Artcltd Lnk Chain Nt Rllr Chain, Iron Or Steel yes 731519 Parts Of Articulated Link Chain Of Iron Or Steel yes 731520 Skid Chain Of Iron Or Steel yes 731581 Stud Link Chain Of Iron Or Steel yes 731582 Chain Nesoi, Welded Link Of Iron Or Steel yes 731589 Chain Of Iron Or Steel Nesoi yes 731590 Parts Of Irn/stl Chain Nesoi yes 52 6-Digit HS Code Product Description Effected by BAA 731600 Anchors, Grapnels And Parts Thereof, Of Iron/steel yes 731700 Nails, Tacks, Drawing Pins Etc Of Iron Or Steel yes 731811 Coach Screws, Threaded, Of Iron Or Steel yes 731812 Other Wood Screws Of Iron Or Steel, Threaded yes 731813 Screw Hooks And Screw Rings Of Iron Or Steel yes 731814 Self-tapping Screws Of Iron Or Steel yes 731815 Threaded Screws And Bolts Nesoi Of Iron Or Steel yes 731816 Nuts Of Iron Or Steel yes 731819 Threaded Articles Of Iron Or Steel, Nesoi yes 731821 Spring Washers And Oth Lock Washers, Iron Or Steel yes 731822 Washers Other Than Lock Washers, Iron Or Steel yes 731823 Rivets Of Iron Or Steel yes 731824 Cotters And Cotter Pins, Of Iron Or Steel yes 731829 Oth Non-threaded Articles (fastnrs) Irn/stl Nesoi yes 731910 Sewing, Darning Or Embroidery Needles, Iron Or Stl yes 731920 Safety Pins Of Iron Or Steel yes 731930 Pins Of Iron Or Steel Nesoi yes 731990 Knitting Needles And Similar Articles, Irn/stl Nes yes 732111 Cooking Appliances Etc For Gas Fuel, Iron Or Steel yes 732112 Cookng Applncs And Plt Wrmrs, Irn/stl For Liq Fuel yes 732113 Cooking Appliances Etc For Solid Fuel, Ir Or Steel yes 732181 Nonelc Dom Appl Nesoi Ios Gas Or Gas A Oth Fuel yes 732182 Nonelc Dom Appl Nesoi Ios Liquid Fuel yes 732183 Nonelc Dom Appl Nesoi Ios Solid Fuel yes 732190 Parts Of Nonelec Domest Cooking Appl, Iron & Steel yes 732211 Radiators For Centrl Htng And Parts, Cast Iron yes 732219 Radiators For Cntrl Htng And Parts, Ios Exc Cstirn yes 732290 Air Htrs A Hot Air Dist Nt Elec Htd Wfan, Prts Ios yes 732310 Ios Wool, Scouring Pads, Gloves Etc, Iron Or Steel yes 732391 Tbl Ktchn Oth hh Artcls A Pts Cst Irn Nt Enmld yes 732392 Oth Tbl, Kitch, Hshld Artic, Enam Cst Irn A Parts yes 732393 Table, Kitchen Etc Articles & Pts, Stainless Steel yes 732394 HH Artcs A Pts Nesoi Enmld Irn Nt Cst Stl Nt Stnls yes 732399 HH Artcs/pts Nesoi Nt Enmld Irn Nt Cst Stl Nt Stls yes 732410 Sinks And Wash Basins Of Stainless Steel yes 732421 Cast Iron Baths Enameled Or Not yes 732429 Baths Of Iron Or Steel, Other Than Cast Iron yes 732490 Other Sanitary Ware, Including Parts, Irn/st Nesoi yes 732510 Cast Articles Nesoi Of Nonmalleable Cast Iron yes 732591 Grinding Balls A Sim Artic For Mills, Cst, Ios Nes yes 732599 Cast Articles Of Iron Or Steel Nesoi yes 760110 Unwrought Aluminum, Not Alloyed no 760120 Unwrought Aluminum Alloys no 760200 Aluminum Waste And Scrap no 760310 Aluminum Powders Of Non-lamellar Structure no 760320 Aluminum Powders Of Lamellar Structure, Flakes no 760410 Aluminum Bars, Rods And Profiles, Not Alloyed no 53 6-Digit HS Code Product Description Effected by BAA 760421 Aluminum Alloy Hollow Profiles no 760429 Aluminum Alloy Bars, Rods And Nonhollow Profiles no 760511 Aluminum Nonalloy Wire, Over 7mm Crsect Max Dimen no 760519 Al Wir Nt Aly Of Whi Th Max C-s Dim Is 7mm Or Less no 760521 Aluminum Alloy Wire, Over 7mm Cross Sect Max Dimen no 760529 Al Wir Nt Aly Of Whi Th Max C-s Dim Is 7mm Or Lss no 760611 Aluminum Nonalloy Rect Plates Etc, Over .2mm Thick no 760612 Aluminum Alloy Rect Plates Etc, Over .2 mm Thick no 760691 Aluminum Nonalloy Plates Etc, Ov .2mm Thick, Nesoi no 760692 Aluminum Alloy Plates Etc, Over .2 mm Thick, Nesoi no 760711 Aluminum Foil, Nov .2mm Th, No Back, Rolled Only no 760719 Aluminum Foil Not Backed Not Ovr .2mm Thck, Nesoi no 760720 Aluminum Foil Not Over 0.2MM Thick, Backed no 760810 Aluminum Nonalloy Tubes And Pipes no 760820 Aluminum Alloy Tubes And Pipes no 760900 Aluminum Tube Or Pipe Fittings no 761010 Alu Dor Win And Their Fra And Thres For Doors no 761090 Aluminum Structures And Parts, Nesoi no 761100 Tanks Etc, Over 300 Liter Capacity, Aluminum no 761210 Aluminum Collapsible Tubular Containers Nt Ov 300l no 761290 Casks Etc, Not Over 300 Liter Cap Nesoi, Aluminum no 761300 Aluminum Containers For Compressed Or Liquefid Gas no 761410 Stranded Wire Etc Of Aluminum With Steel Core no 761490 Stranded Wire Etc, No Elec Insul Nesoi, Aluminum no 761511 Alum Pot Scours, Scourng/polishng Pads/gloves, Etc no 761519 Table,kitcen,& Other Household Articles, Aluminum no 761520 Aluminum Sanitary Ware And Parts Thereof no 761610 Nails, Tacks, Staples, Screws, Nuts Etc, Aluminum no 761691 Cloth, Grill, Netting And Fencing Of Aluminum Wire no 761699 Articles Of Aluminum, N.E.S.O.I. no 54 Bibliography Adina, C., Gabriela, C., Roxana-Denisa, S. (2015). Country-of-Origin Effects on Perceived Brand Positioning. Procedia Economics and Finance, 23, pp 422-427. Atkin, D. (2013). Trades, Tastes, and Nutrition in India. American Economic Review, 103(5), pp 342-352. Beattie, A. (2018). Spillover Effect. Investopedia. Baughn C.C., Yaprak, A. (1996). Economic Nationalism: Conceptual and Empirical Development. Political Psychology, 17(4), pp 759-778 Chao, P. (1998). Impact of Country-of-Origin Dimensions on Product Quality and Design Quality Perceptions. Journal of Business Research, 42(1), pp 1-6. Costa, C., Jorge, C., Goldszmidt, R. (2016). A contingent approach to country-of-origin effects on foreign products evaluation: Interaction of facets of country image with product classes. International Business Review, 25(5), pp 1066-1075. Darby, K, Batte, M.T., Ernst, S., Roe, B. (2006). Willingness to Pay for Locally Produced foods: A Customer Intercept Study of Direct Market and Grocery Store Shoppers. American Agricultural Economics Association, pp 1-31. Dixon, P. B., Rimmer, Maureen T., and Waschik, Robert G. Macro, Industry and Regional Effects of Buy America(n) Programs: USAGE Simulations. Centre of Policy Studies Working Paper No. G-271. Melbourne: Victoria University, 2017. Donald J. Trump for President, Inc. (2019). “Economy And Jobs. “ Promises Kept, https://www.promiseskept.com/achievement/overview/economy-and-jobs/. Frank, D. Buy American, The Untold Story of Economic Nationalism. Boston: Beacon Press, 1999. Fisher, M. Over four decades, Trump’s one solid stance: A hard line on trade. Washington Post, 2018. Hamazaoui-Essoussi, L., Merunka, D., Bartikowski, B. (2011). Brand origin and country of manufacture influences on brand equity and the moderating role of brand typicality. Journal of Business Research, 64(9), pp 973-978. Helleiner, E. (2002). Economic Nationalism as a Challenge to Economic Liberalism? Lessons from the 19th Century. International Studies Quarterly, 46(3), pp 307- 329. 55 Heller, L. (2016). Walmart’s Made in The USA Claim: Fact or Fiction? Forbes Media LLC. Hufbauer, G. C. and Jung, E. “‘Buy American’ Is Bad for Taxpayers and Worse for Exports.” Peterson Institute for International Economics, 2017. Hoffman, V. and Gatobu, K. (2014). Growing their own: Unobservable quality and the value of self-provisioning. Journal of Development Economics, 106, pp 168- 178. Hu, W., Batte, M.T., Woods, T., Ernst, S. (2012). Growing their Own: Unobservable Quality and the Value of Self-provisioning. Jounral of Developmental Economics, 106, pp 168-178. Irwin, D.A. (1996) The Smoot-Hawley Tariff: A Quantitative Assessment. Review of Economics and Statistics, 80(2), pp 326-334. Ittersum, K.V., Candel, J.J.M., Meulenberg, M.T.G. (2003). The influence of the image of a product’s region of origin on product evaluation. Journal of Business Research, 56(3), pp Johnson, H.G. (1965). A Theoretical Model of Economic Nationalism in New and Developing States. Political Science Quarterly, 80(2), pp 169-185. Lee, J.K., Lee, B.K., Lee, W.N. (2013). Country-of-origin fit’s effect on consumer product evaluation in cross-border strategic brand alliance. Journal of Business Research, 66(3), pp 354-363. Lin, C., He, X. (2017). Expanding near the home base or venture far? The influence of home country state on the economic distance of foreign direct investments. Journal of Business Research, 75, pp 95-107. Manuel, K.A., A.M.Dolan, B.J. Murrill, R.M. Perry and S.P. Mulligan (2016), “Domestic content restrictions: the Buy American Act and complementary provisions in federal law”, Congressional research Services, 7-5700, R43354, Morey, M. (2016). Preferences and the home bias in trade. Journal of Development Economics, 121, pp. 24-37. Pan, S., Pezzuti, P., Pechmann, C. (2018). Hyperopia and frugality: Different motivational drivers and yet similar effects on consumer spending. Journal of Business Research. Pappu, R. Quester, P., Cooksey, R. (2006). Consumer-based brand equity and country- of-origin relationship: Some empirical evidence. European Journal of Marketing, 40(5/6), pp 696-717. 56 Phillip, A. Trump signs ‘Buy American, Hire American’ executive order, promising to fight for American workers. The Washington Post, 2017. Sahoo, A., Shirmali, G. (2013). The effectiveness of domestic content criteria in India’s Solar Mission. Energy Policy, 62, pp 1470-1480. Trefler, D. (1995). The Case of the Missing Trade and other Mysteries. American Economic Review, 95(5), pp 1029-1046. U.S. Congress, OFfice of Technology Assessment, Advanced Materials by Design, OTA-#-351. Washington, DC: U.S. Government Printing Office, 1968. Veale, R., Quester, P. (2009). Do consumer expectations match experience? Predicting the influence of price and country of origin on perceptions of product quality. International Business Review, 18(2), pp 134-144. The White House. Presidential Executive Order on Buy American and Hire American. April 18, 2017. Whiting, T. “‘Buy American’ Laws: A Costly Policy Mistake That Hurts Americans.” The Heritage Foundation, 2017. 57