Trade,Foreign Direct Investment,and Immigration Policy Making 815 The article proceeds as follows.I begin by theorizing how trade openness and increased firm mobility can lead to restrictions on immigration because of their effects on firms.Next,I discuss the ways in which tariff policy and the ability to invest overseas were out of the control of US senators.Third,I provide evidence that trade policy and firm mobility help explain this variation,and I examine the evi- dence for the alternative explanations.Fourth,I provide evidence showing how firm lobbying on immigration has changed with trade openness and increases in firm mobility.Finally,I conclude by discussing the implications of this article for research on other economic policies. Trade Policy,Firm Mobility,and Support for Immigration In this section,I examine how changes in trade policy and firm mobility affect firms' support for openness to low-skill immigrants and how this,in turn,affects policy- makers'support for open LSIP.I argue that immigration policy toward low-skill workers and the policy toward high-skill workers should be studied separately; although,they often get bundled together conceptually and in legislation.16 These policies target two different populations that are not interchangeable.Low-skill immi- grants do not have the skills to take the place of high-skill immigrants and high-skill immigrants are unlikely to give up the wage premium that their skills provide them by taking low-skill jobs.Policy similarly is often used to target these populations;many countries have policies specifically targeting high-skill or low-skill workers.In general,when states are open to low-skill migrants they are also open to high-skill migrants but not vice-versa.17 Further,the flow of low-skill migrants is more politicized than the flow of high- skill migrants.Natives tend to have much more favorable views of high-skill immi- grants than low-skill immigration.18 Nor do these preferences seem to be a new feature of politics.In the United States,for example,most of the anti-immigrant senti- ment has been targeted toward the least-skilled migrant group:the Irish in the 1840s, the Chinese in the mid-1800s,the southern and eastern Europeans at the turn of the last century,and Hispanics today.It is,therefore,politically easier for a politician to support a more open policy toward high-skill immigrants than to low-skill immigration. Although beyond the scope of this research,the politics of high-skill immigration may also be affected by trade openness and firm-mobility as many high-skill occu- pations can be easily outsourced overseas.However,there is less opposition to 16.These policies are bundled together in one piece of legislation for the same reasons that there is omnibus legislation in other policy areas:fragmentation across committees,divided government,and easy minority obstruction.Krutz 2001. 17.Peters forthcoming. 18.For example,see Goldstein and Peters 2014;and Hainmueller and Hiscox 2010. 19.Blinder 2007
The article proceeds as follows. I begin by theorizing how trade openness and increased firm mobility can lead to restrictions on immigration because of their effects on firms. Next, I discuss the ways in which tariff policy and the ability to invest overseas were out of the control of US senators. Third, I provide evidence that trade policy and firm mobility help explain this variation, and I examine the evidence for the alternative explanations. Fourth, I provide evidence showing how firm lobbying on immigration has changed with trade openness and increases in firm mobility. Finally, I conclude by discussing the implications of this article for research on other economic policies. Trade Policy, Firm Mobility, and Support for Immigration In this section, I examine how changes in trade policy and firm mobility affect firms’ support for openness to low-skill immigrants and how this, in turn, affects policymakers’ support for open LSIP. I argue that immigration policy toward low-skill workers and the policy toward high-skill workers should be studied separately; although, they often get bundled together conceptually and in legislation.16 These policies target two different populations that are not interchangeable. Low-skill immigrants do not have the skills to take the place of high-skill immigrants and high-skill immigrants are unlikely to give up the wage premium that their skills provide them by taking low-skill jobs. Policy similarly is often used to target these populations; many countries have policies specifically targeting high-skill or low-skill workers. In general, when states are open to low-skill migrants they are also open to high-skill migrants but not vice-versa.17 Further, the flow of low-skill migrants is more politicized than the flow of highskill migrants. Natives tend to have much more favorable views of high-skill immigrants than low-skill immigration.18 Nor do these preferences seem to be a new feature of politics. In the United States, for example, most of the anti-immigrant sentiment has been targeted toward the least-skilled migrant group: the Irish in the 1840s, the Chinese in the mid-1800s, the southern and eastern Europeans at the turn of the last century, and Hispanics today. It is, therefore, politically easier for a politician to support a more open policy toward high-skill immigrants than to low-skill immigration. Although beyond the scope of this research, the politics of high-skill immigration may also be affected by trade openness and firm-mobility as many high-skill occupations can be easily outsourced overseas.19 However, there is less opposition to 16. These policies are bundled together in one piece of legislation for the same reasons that there is omnibus legislation in other policy areas: fragmentation across committees, divided government, and easy minority obstruction. Krutz 2001. 17. Peters forthcoming. 18. For example, see Goldstein and Peters 2014; and Hainmueller and Hiscox 2010. 19. Blinder 2007. Trade, Foreign Direct Investment, and Immigration Policy Making 815
816 International Organization high-skill immigration and governments may favor high-skill immigration given these immigrants'contributions to the tax base.20 Therefore,even if there is less support for high-skill immigration from firms,high-skill immigration may remain open. Finally,whereas high-skill immigration gains more attention from the media,low- skill immigration is,arguably,more important.The vast majority of all potential immigrants have been low-skill workers.21 Moreover,studies have shown that the migrants'remittances lead to better outcomes for their families and communities and that increased low-skill migration would greatly increase both developing nations'and world income.22 Given the benefits of low-skill migration,IPE scholars should be interested in why low-skill immigration is so restricted in most wealthy states,including the United States,today. Although I examine US policy toward low-skill immigration in the empirical section,my argument applies to any low-skill labor scarce state;although,these states'policies may not "look"like US LSIP.23 First,there are different ways to restrict LSIP.For example,the United States uses numerical quotas to control the flow of low-skill immigrants whereas Canada and Australia,among others,use a point system.The United States,Canada,and Australia adopted these regulations in the 1960s and 1970s with the same goal in mind;namely,they wanted to replace their racist,national origin restrictions but still restrict low-skill,nonwhite immigration.4 Although these states chose different regulations,they have all been relatively successful at restricting low-skill immigration.25 Nonetheless,these policies have had different effects on the number of high-skill immigrants entering these countries,something that is beyond the scope of this article.Further,as dis- cussed next,the level of openness will depend on states'openness to trade and the mobility of their firms.Thus,it may not be surprising that these states have different LSIPs even though these policies are driven by the same factors. Finally,the argument in this article is,like most,an "all else equal"argument. There are many groups in the polity that could affect LSIP:firms,labor,nativists,tax- payers,and immigrants themselves.In classic economic models,native labor dislikes immigrants because low-skill immigrants compete for jobs and push wages down (which is exactly why firms support immigration).Since recent survey data do not support these economic models,some scholars have turned to cultural reasons for opposition to immigration.26 Nativists dislike immigrants for the obvious reason: immigrants are different from natives.Other scholars have examined the fiscal 20.Medina 2010. 21.See Hatton and Williamson 2005:and United Nations Development Program 2009. 22.See Brown 2006;and Hatton and Williamson 2005. 23.See Peters forthcoming,for a discussion of how this argument applies to other low-skill labor scarce states. 24.See Jupp 2002;Kelley and Trebilcock 1998;and Zolberg 2006. 25.Peters forthcoming. 26.For example,Hainmueller and Hiscox 2010
high-skill immigration and governments may favor high-skill immigration given these immigrants’ contributions to the tax base.20 Therefore, even if there is less support for high-skill immigration from firms, high-skill immigration may remain open. Finally, whereas high-skill immigration gains more attention from the media, lowskill immigration is, arguably, more important. The vast majority of all potential immigrants have been low-skill workers.21 Moreover, studies have shown that the migrants’ remittances lead to better outcomes for their families and communities and that increased low-skill migration would greatly increase both developing nations’ and world income.22 Given the benefits of low-skill migration, IPE scholars should be interested in why low-skill immigration is so restricted in most wealthy states, including the United States, today. Although I examine US policy toward low-skill immigration in the empirical section, my argument applies to any low-skill labor scarce state; although, these states’ policies may not “look” like US LSIP.23 First, there are different ways to restrict LSIP. For example, the United States uses numerical quotas to control the flow of low-skill immigrants whereas Canada and Australia, among others, use a point system. The United States, Canada, and Australia adopted these regulations in the 1960s and 1970s with the same goal in mind; namely, they wanted to replace their racist, national origin restrictions but still restrict low-skill, nonwhite immigration.24 Although these states chose different regulations, they have all been relatively successful at restricting low-skill immigration.25 Nonetheless, these policies have had different effects on the number of high-skill immigrants entering these countries, something that is beyond the scope of this article. Further, as discussed next, the level of openness will depend on states’ openness to trade and the mobility of their firms. Thus, it may not be surprising that these states have different LSIPs even though these policies are driven by the same factors. Finally, the argument in this article is, like most, an “all else equal” argument. There are many groups in the polity that could affect LSIP: firms, labor, nativists, taxpayers, and immigrants themselves. In classic economic models, native labor dislikes immigrants because low-skill immigrants compete for jobs and push wages down (which is exactly why firms support immigration). Since recent survey data do not support these economic models, some scholars have turned to cultural reasons for opposition to immigration.26 Nativists dislike immigrants for the obvious reason: immigrants are different from natives. Other scholars have examined the fiscal 20. Medina 2010. 21. See Hatton and Williamson 2005; and United Nations Development Program 2009. 22. See Brown 2006; and Hatton and Williamson 2005. 23. See Peters forthcoming, for a discussion of how this argument applies to other low-skill labor scarce states. 24. See Jupp 2002; Kelley and Trebilcock 1998; and Zolberg 2006. 25. Peters forthcoming. 26. For example, Hainmueller and Hiscox 2010. 816 International Organization
Trade,Foreign Direct Investment,and Immigration Policy Making 817 costs of immigrants as the source of anti-immigrant sentiment.27 Finally,immigrants typically are pro-immigration to protect their position in society and/or to bring in friends and family;however,immigrants tend not to be a powerful group.Until they gain the rights of citizens (and,sometimes,not even then),immigrants can be expelled from the country,which limits their political power.Nonetheless,the rela- tive power of these groups is likely to affect the level of openness to immigration and these groups are likely to have different amounts of power in different countries, which would affect their openness to immigration.In the empirical analysis below, I control for these other factors. I focus on firms'support for immigration because whereas these other groups likely play a role in LSIP formation,firms are the most powerful group that could be pro-immi- gration.Given firms'important role in LSIP,I examine how their preferences change because of changes in productivity,trade openness,and firm mobility and how policy- makers respond to these changes.In the United States,a low-skill labor-scarce economy,trade openness,and increases in firm mobility should affect the preferences of firms that are low-skill labor intensive and/or less productive the most,which also are the firms that use the most low-skill immigrant labor.Openness to foreign goods and increasing firm mobility,therefore,may not need to affect the entire economy for them to have an effect on LSIP;instead,they need to affect only low-skill-intensive firms. Low-Skill Immigration Policy Under Autarky To contrast the effects of trade openness and firm mobility,I begin by examining support for LSIP by firms under trade protection and limited opportunities to move abroad (autarky).The goal of firms when they interact with policy-makers is to convince policy-makers to pass policies that decrease their costs and increase their profitability.Firms do not necessarily have a preference over which policy policy-makers choose so they are likely indifferent between LSIP that lowers their labor costs and some other policy that lowers other costs or increases profits. Firms across and within industries differ in their need for low-skill labor;some indus- tries are more capital and/or high-skill-intensive and some firms within an industry are more productive.The terms low-skill labor intensive and low-productivity are used somewhat interchangeably.In economics,these two aspects of the firms are treated as separate and the effect of trade is modeled differently:differences in the factor intensity of production is modeled by the Ricardo-Viner model and productivity differences are modeled by the Melitz and other similar models.28 In the real world,firms differ on both dimensions-some firms in low-skill labor-intensive industries are more productive than others and some industries are more capital/high-skill-intensive than others. What is important for this analysis is that these two dimensions collapse when one examines preferences on immigration;henceforth,less-productive and 27.For example,see Gimpel and Edwards 1999;and Hanson,Scheve,and Slaughter 2007. 28.Melitz 2003
costs of immigrants as the source of anti-immigrant sentiment.27 Finally, immigrants typically are pro-immigration to protect their position in society and/or to bring in friends and family; however, immigrants tend not to be a powerful group. Until they gain the rights of citizens (and, sometimes, not even then), immigrants can be expelled from the country, which limits their political power. Nonetheless, the relative power of these groups is likely to affect the level of openness to immigration and these groups are likely to have different amounts of power in different countries, which would affect their openness to immigration. In the empirical analysis below, I control for these other factors. I focus on firms’ support for immigration because whereas these other groups likely play a role in LSIP formation, firms are the most powerful group that could be pro-immigration. Given firms’ important role in LSIP, I examine how their preferences change because of changes in productivity, trade openness, and firm mobility and how policymakers respond to these changes. In the United States, a low-skill labor-scarce economy, trade openness, and increases in firm mobility should affect the preferences of firms that are low-skill labor intensive and/or less productive the most, which also are the firms that use the most low-skill immigrant labor. Openness to foreign goods and increasing firm mobility, therefore, may not need to affect the entire economy for them to have an effect on LSIP; instead, they need to affect only low-skill-intensive firms. Low-Skill Immigration Policy Under Autarky To contrast the effects of trade openness and firm mobility, I begin by examining support for LSIP by firms under trade protection and limited opportunities to move abroad (autarky). The goal of firms when they interact with policy-makers is to convince policy-makers to pass policies that decrease their costs and increase their profitability. Firms do not necessarily have a preference over which policy policy-makers choose so they are likely indifferent between LSIP that lowers their labor costs and some other policy that lowers other costs or increases profits. Firms across and within industries differ in their need for low-skill labor; some industries are more capital and/or high-skill-intensive and some firms within an industry are more productive. The terms low-skill labor intensive and low-productivity are used somewhat interchangeably. In economics, these two aspects of the firms are treated as separate and the effect of trade is modeled differently: differences in the factor intensity of production is modeled by the Ricardo-Viner model and productivity differences are modeled by the Melitz and other similar models.28 In the real world, firms differ on both dimensions—some firms in low-skill labor-intensive industries are more productive than others and some industries are more capital/ high-skill-intensive than others. What is important for this analysis is that these two dimensions collapse when one examines preferences on immigration; henceforth, less-productive and 27. For example, see Gimpel and Edwards 1999; and Hanson, Scheve, and Slaughter 2007. 28. Melitz 2003. Trade, Foreign Direct Investment, and Immigration Policy Making 817
818 Interational Organization low-skill-intensive firms and capital/high-skill and more productive firms will be referred to as low-skill-and high-skill-intensive,respectively.Low-skill-intensive firms benefit from open LSIP more than firms that are high-skill-intensive.29 Assuming that each firm's political capital is limited,low-skill-intensive firms should be willing to spend more political capital on LSIP than high-skill-intensive firms.It is not necessarily the case that high-skill-intensive firms do not want more low-skill immi- gration;it is simply that they prefer to spend their political capital elsewhere,including on high-skill immigration policy.As the proportion of high-skill-intensive firms increases,policy-makers will receive less political capital for the same level of LSIP and given the existence of groups that oppose immigration,one should expect that, all else equal,senators from states with higher high-skill intensity will support more restrictive LSIP.If one allows for oligopoly,increasing immigration conveys an advan- tage to firms that use more labor because it lowers these firms'costs to a greater extent, allowing them to capture a greater share of the market.High-skill-intensive firms may want to keep their competitive edge by giving contributions for restrictions. Trade closure,all else equal,should have an effect on the composition of firms in the United States;it should increase the number and/or size of low-skill-intensive firms,as,under the Ricardo-Viner model,domestic production becomes more com- petitive or,under the Melitz model,less-productive firms are able to stay in business.30 Thus,trade restrictions increase the demand for low-skill labor,increase the wages firms pay and,without an increase in the labor force,may erase gains from trade protection.31 Further,they increase the wage for low-skill labor in the nontrad- able or export sector as well.Therefore,one should expect that higher trade barriers should lead to increased support by firms and senators for open LSIP.Similarly,the inability to move production overseas (or low firm mobility)will also increase the pro- duction of low-skill labor-intensive goods,as firms have no choice but to produce at home,and increase support by firms and by senators for more open LSIP.32 Low-Skill Immigration Policy Under Open Trade and Low Firm Mobility Under autarky,firm preferences-and their willingness to spend political capital- are driven by their skill intensity.Opening trade does not affect these preferences; instead,it increases the proportion and/or size of firms that are high-skill-intensive. 29.Helpman,Itskhoki,and Redding 2009 argue that less productive firms will employ more low-skilled workers than more productive firms. 30.Melitz 2003. 31.Helpman,Itskhoki,and Redding 2009. 32.If trade is re-restricted and/or firm mobility is limited,there should be an increase in low-skill-inten- sive production,with a concomitant increase in demand for open LSIP.Yet trade and capital restrictions often are enacted during recessions.One might not expect firms to expand production during a recession and if they do expand,they can use native,previously unemployed labor until unemployment returned to its natural rate.At that point,one would expect wages to rise and firms to pressure the govemment for open LSIP
low-skill-intensive firms and capital/high-skill and more productive firms will be referred to as low-skill- and high-skill-intensive, respectively. Low-skill-intensive firms benefit from open LSIP more than firms that are high-skill-intensive.29 Assuming that each firm’s political capital is limited, low-skill-intensive firms should be willing to spend more political capital on LSIP than high-skill-intensive firms. It is not necessarily the case that high-skill-intensive firms do not want more low-skill immigration; it is simply that they prefer to spend their political capital elsewhere, including on high-skill immigration policy. As the proportion of high-skill-intensive firms increases, policy-makers will receive less political capital for the same level of LSIP and given the existence of groups that oppose immigration, one should expect that, all else equal, senators from states with higher high-skill intensity will support more restrictive LSIP. If one allows for oligopoly, increasing immigration conveys an advantage to firms that use more labor because it lowers these firms’ costs to a greater extent, allowing them to capture a greater share of the market. High-skill-intensive firms may want to keep their competitive edge by giving contributions for restrictions. Trade closure, all else equal, should have an effect on the composition of firms in the United States; it should increase the number and/or size of low-skill-intensive firms, as, under the Ricardo-Viner model, domestic production becomes more competitive or, under the Melitz model, less-productive firms are able to stay in business.30 Thus, trade restrictions increase the demand for low-skill labor, increase the wages firms pay and, without an increase in the labor force, may erase gains from trade protection.31 Further, they increase the wage for low-skill labor in the nontradable or export sector as well. Therefore, one should expect that higher trade barriers should lead to increased support by firms and senators for open LSIP. Similarly, the inability to move production overseas (or low firm mobility) will also increase the production of low-skill labor-intensive goods, as firms have no choice but to produce at home, and increase support by firms and by senators for more open LSIP.32 Low-Skill Immigration Policy Under Open Trade and Low Firm Mobility Under autarky, firm preferences—and their willingness to spend political capital— are driven by their skill intensity. Opening trade does not affect these preferences; instead, it increases the proportion and/or size of firms that are high-skill-intensive. 29. Helpman, Itskhoki, and Redding 2009 argue that less productive firms will employ more low-skilled workers than more productive firms. 30. Melitz 2003. 31. Helpman, Itskhoki, and Redding 2009. 32. If trade is re-restricted and/or firm mobility is limited, there should be an increase in low-skill-intensive production, with a concomitant increase in demand for open LSIP. Yet trade and capital restrictions often are enacted during recessions. One might not expect firms to expand production during a recession and if they do expand, they can use native, previously unemployed labor until unemployment returned to its natural rate. At that point, one would expect wages to rise and firms to pressure the government for open LSIP. 818 International Organization
Trade,Foreign Direct Investment,and Immigration Policy Making 819 I begin by assuming that trade openness is exogenous to both the firm and the policy- maker,which I relax later.The effect of trade openness does not depend on whether the trade is opened to countries with different endowments(Ricardo-Viner model)or the same endowments (Melitz model).Under the Ricardo-Viner model,trade open- ness affects firms by increasing the price of high-skill-intensive goods and decreasing the price of low-skill-intensive goods.As prices for low-skill-intensive goods decrease,firms that produce these goods (the threatened firms)have to decrease costs or close.Under models of intraindustry trade (the Melitz model),only the most productive firms can export.As the highly productive firms export,they increase the amount of labor they need and bid up the real wages,forcing the least productive firms to close.33 Under either model,low-skill-intensive firms face increased competition because of trade openness and are likely to close.34 If low- skill-intensive firms close,they will not provide any support for open LSIP and sena- tors should vote for restrictions more often.Further,the closure of some firms may lead others to spend less political capital on LSIP as well.When firms close,they lay off their workers which leads to lower wages,negating the need for as much low-skill immigration,even for nontradable industries.Given that firms have many issues on which they may want to spend their political capital,a decrease in the wage because of other firms'layoffs should make them less likely to spend political capital on LSIP. Because closing is an undesirable outcome for the firm,firms likely will pursue strategies to stay in business,including increasing their use of high-skill labor.As I have shown,if firms increase their use of high-skill labor,they will be less suppor- tive of open LSIP and senators will vote for restrictions more often.Firms have long understood that increasing their use of technology will decrease their need for low- skill labor and make them more competitive.For example,in his address to the New England Cotton Manufacturers'Association in 1896,Mayor Josiah Quincy of Boston argued that "As other sections,nearer to the sources of the supply of fuel, enabled to command cheaper labor in some respects,coming to the market and taking up manufacturing industries as the South is largely doing,it seems to me that those who are engaged in similar industries in New England must recognize the fact that in order to hold their supremacy they must put more brains,more skill and more education in the carrying on of the manufacturing business."35 In this case,the competition was not only from overseas but also domestic competition from the South,from which it was impossible to gain any trade protection. 33.Melitz2003,1716. 34.These results assume that production by high-skill-intensive firms does not increase so much as to increase the economy-wide demand for low-skill labor.The expanding firms use much more high-skill labor than low-skill labor,the firms that exit the economy,on the other hand,release much native low- skill labor.It is likely that the laid-off workers will more than meet the demand for low-skill labor. Empirically,rising wage inequality because of increases in productivity and trade seem to bear this out. For example,Feenstra and Hanson 1996. 35.Quincy1896,66
I begin by assuming that trade openness is exogenous to both the firm and the policymaker, which I relax later. The effect of trade openness does not depend on whether the trade is opened to countries with different endowments (Ricardo-Viner model) or the same endowments (Melitz model). Under the Ricardo-Viner model, trade openness affects firms by increasing the price of high-skill-intensive goods and decreasing the price of low-skill-intensive goods. As prices for low-skill-intensive goods decrease, firms that produce these goods (the threatened firms) have to decrease costs or close. Under models of intraindustry trade (the Melitz model), only the most productive firms can export. As the highly productive firms export, they increase the amount of labor they need and bid up the real wages, forcing the least productive firms to close.33 Under either model, low-skill-intensive firms face increased competition because of trade openness and are likely to close.34 If lowskill-intensive firms close, they will not provide any support for open LSIP and senators should vote for restrictions more often. Further, the closure of some firms may lead others to spend less political capital on LSIP as well. When firms close, they lay off their workers which leads to lower wages, negating the need for as much low-skill immigration, even for nontradable industries. Given that firms have many issues on which they may want to spend their political capital, a decrease in the wage because of other firms’ layoffs should make them less likely to spend political capital on LSIP. Because closing is an undesirable outcome for the firm, firms likely will pursue strategies to stay in business, including increasing their use of high-skill labor. As I have shown, if firms increase their use of high-skill labor, they will be less supportive of open LSIP and senators will vote for restrictions more often. Firms have long understood that increasing their use of technology will decrease their need for lowskill labor and make them more competitive. For example, in his address to the New England Cotton Manufacturers’ Association in 1896, Mayor Josiah Quincy of Boston argued that “As other sections, nearer to the sources of the supply of fuel, enabled to command cheaper labor in some respects, coming to the market and taking up manufacturing industries as the South is largely doing, it seems to me that those who are engaged in similar industries in New England must recognize the fact that in order to hold their supremacy they must put more brains, more skill and more education in the carrying on of the manufacturing business.”35 In this case, the competition was not only from overseas but also domestic competition from the South, from which it was impossible to gain any trade protection. 33. Melitz 2003, 1716. 34. These results assume that production by high-skill-intensive firms does not increase so much as to increase the economy-wide demand for low-skill labor. The expanding firms use much more high-skill labor than low-skill labor; the firms that exit the economy, on the other hand, release much native lowskill labor. It is likely that the laid-off workers will more than meet the demand for low-skill labor. Empirically, rising wage inequality because of increases in productivity and trade seem to bear this out. For example, Feenstra and Hanson 1996. 35. Quincy 1896, 66. Trade, Foreign Direct Investment, and Immigration Policy Making 819