Integration and Disintegration: Trade and Labor Market Integration

After World War II, the victors—the U


I Introduction
In the last few years, it has seemed as if the forces of disintegration have reigned supreme. First, there was the Brexit vote in Summer 2016, followed by the election of Donald Trump as the president of the United States in the fall of that year, and the increasing fortunes of far-right parties and politicians throughout Europe. While there has been some "buyer's remorse" by publics throughout the world-Brexit was pushed back several times, the Democratic party in the U.S. was resurgent in the 2018 congressional elections, and center and left parties have been able to fend off challenges by the far-right across Europe and even gained seats in the European Parliament-, it is clear that there is growing public discontent with globalization, the greater integration of economic markets for goods, services, and money.
What has been most notable about this anger is the high level of resentment towards immigrants, even though immigration has not been part of the post-war liberal international order (LIO). That immigration was left out of the LIO was intentional. The changes of the late nineteenth century and the Interwar Era had increased the power of native labor and with it the need to protect domestic workers from the forces of the international market. 1 When creating the new institutions that would help maintain exchange rates and moderate capital flows (the International Monetary Fund or IMF ) and increase trade flows (the General Agreement on Tariffs and Trade, later the World Trade Organization, or GATT/WTO), the American and British framers had to allow governments to manage their domestic labor markets. Any supra-national organization or agreements on the movement of labor that threatened domestic labor markets were out of the question. 2 5 of pro-immigration supporter if she moves to the right on immigration-because those groups no longer care about the issue-she can easily move to capture additional votes. There have almost always been groups and voters who oppose immigration; if they are not counterbalanced by powerful pro-immigration groups, politicians restrict immigration to obtain those votes. By moving to the right, they legitimize anti-immigrant views and bring them into the debate. This then allows political entrepreneurs, like Nigel Farange of the U.K.
Independence Party or Donald Trump, to move from the fringe into the mainstream. In turn, the increased political space and power of anti-immigrant forces then leads to greater restrictions on immigration. Thus, the decrease in business support has emboldened antiimmigrant forces and led to further restrictions on immigration. 55 I test the implications of my argument empirically using two different types of data.
First, I use data on trade policy and immigration policy in the West (Canada, Australia, New Zealand, the US, and Europe) since World War II to show that immigration and trade are inversely related. When trade has been more restricted, immigration has been more open, as policymakers cater to business demands for greater openness. Yet, when trade opens, immigration becomes restricted as businesses close allowing policymakers to cater to anti-immigrant groups. The ability of firms to move production, as measured by greater capital openness, only magnifies this effect. Second, I dive deeper into the relationship between trade and immigration using senate roll call votes on immigration. I find that when trade openness leads firms in the senator's home state to close, the senator is more likely to vote for restrictions on immigration. Similarly, when more firms have the ability to move or when they mechanize production, senators too are more likely to vote for 6 restrictions. Together, the change in firm support helps explain why Republicans are increasingly anti-immigration.
The way that greater integration of world's economies was achieved then has sowed the seeds for its backlash. By integrating trade and capital markets first, the framers of the LIO forestalled the integration of labor markets. In the short run, this helped to protect native labor but in the long-run it has had the opposite effect: instead of competing against foreign workers in the more regulated labor markets of the Global North, native workers have had to compete against them in the much-less-regulated labor markets of the Global South. This has likely led to more job losses for native workers in the Global North than would have been the case if states had allowed more immigration. Further, it decreased the chance for meaningful contact between natives and immigrants that might prevent antiimmigrant sentiment. Scholars going back to Allport have noted that high quality contact decreases prejudice. 6 The lack of quality contact at work may have allowed fears of immigrants to increase, since few had meaningful experiences with immigrants with which to build counter narratives. This then left many open to emboldened politicians who could frame immigrants as the cause of citizens' problems, instead of more likely culprits like trade and automation. 7 This article continues as follows. First, I detail the links between trade, firm mobility, productivity, and immigration politics and policy. In this section, I draw upon 7 economic theories on trade including both Ricardian comparative advantage and "new" new trade theory to understand how trade, firm mobility, and productivity affect immigration. Then, I discuss how the LIO was put together in the post-war era and what this has meant for immigration, using data on immigration and trade policies in the West.
Third, I discuss how trade and capital mobility freed politicians to move to the right on immigration using a case study of voting in the U.S. senate. Finally, I conclude with implications for integration.

II Theory
I argue that trade openness and firm mobility decrease political support for low-skill immigration in wealthy countries. Although all countries experience some immigration, my argument applies especially to states to which many immigrants want to move-states that are wealthy relative to other states in their region or the world. These states must decide whether and how to control their borders. Throughout this section, I draw upon both economic models of trade and political science models of politics.
I assume that policymakers want to stay in office. In all societies, policymakers must have both elite and mass support to stay in office; the amount of one versus the other changes depending on whether the state is a democracy, in which mass attitudes matter a lot, or autocracy, in which elite attitudes are likely to matter more. 8 When we think about how disintegration has occurred in the recent past, we often think about the change in support for globalization and integration among democracies. Thus, for the rest of the discussion and in presenting the evidence, I focus on democracies and democratic politics.
When policymakers in a democracy craft immigration policy, I assume that they more likely to care about the views of voters the closer it is to an election 12 or when an issue is highly salient to the voters. 13 Farther from an election or when an issue is less salient, politicians are more likely to prioritize the views interest groups or their own ideology.
Regardless, interest groups' positions are likely to factor into politicians' calculations.
Business interests, in particular, are likely to play a large role-businesses provide jobs for constituents and needed tax revenue-in addition to whatever support they may provide for re-election. 14 Thus, as lamented by many, business interests are likely to have a strong effect over policy. This paper, then, considers how businesses' support for low-skill immigration changes with globalization. From here on, I use the term "lobby" to encompass all activities that firms could do that would increase the likelihood that policymakers take their views into account, including providing campaign contributions, working on get out the vote drives, lobbying officials, or advising on policy. I assume that any firm that wants to lobby can do so on its own or as part of a larger organization. 15 Firms are strategic in their lobbying. Most importantly, firms are not going to lobby in excess of the benefit they will get from a policy change. 16 Lobbying is costly in terms of both money and time that could be spent on other activities of the firm or be given as profit to the owners of the firm. If a firm will not benefit from a change in immigration policy, for example if the firm is a manufacturing firm and the proposed change allows in agricultural workers, the firm will be unlikely to lobby for the policy. Lobbying effort will likely also be 10 affected by the probability that the policy can be passed. 17 Firms have different preferences over low-skill immigration depending on the other options they have to produce besides using low-skill labor in their home country. I examine two cleavages that affect business preferences. The first cleavage is based on the low-skill labor intensity of production: firms that use much low-skill labor will favor open low-skill immigration whereas firms that use little low-skill labor will be indifferent. As technology advances, more firms will move from the low-skill-intensive sector to the high-skill-intensive sector, 18 reducing support for low-skill immigration. The second cleavage is mobility across international borders: sectors that are relatively immobile are more likely to support open immigration than sectors that can move overseas. As firm mobility increases, more firms move from the immobile category to the mobile category, again reducing support on low-skill immigration. Trade acts on these preferences by changing the number of firms in the lowskill-intensive, low-mobility sector. Trade restrictions increase the number of firms in this category and allow these firms to grow in size. Trade openness decreases both the number of firms in this category and shrinks the size of existing firms.
The total amount of lobbying on immigration by firms-both the total number of firms lobbying and how much each firm lobbies-is based on the number of firms that use low-skill labor. When low-skill-intensive firms close due to trade competition, when they move overseas, or when they adopt labor-saving technology, support for open immigration drops. This, in turn, allows anti-immigrant forces to have more influence over immigration policy and leads to more immigration restrictions.

A Trade's & international competition's effects on business
In classic trade theory, states trade based on their comparative advantage. Wealthy states are abundant in capital and high-skill labor and export goods that use much of those factors of production in their production (high-skill-intensive goods) whereas developing countries are abundant in low-skill labor and export goods that are produced with low-skill labor (low-skill-intensive goods). When states open their borders to trade, the export sector, in which they have a comparative advantage, will expand as there is greater demand for these goods from overseas and the import sector, in which they are at a comparative disadvantage, will contract as there is more competition for these goods from overseas. 19 For wealthy countries, this means that the low-skill-intensive sectors of the economy will contract.
These are the same sectors that employ low-skill immigrants. When firms in these sectors close, they take their support for low-skill immigration with them. Without businesses' support for immigration, anti-immigration supporters gain relatively more power and policymakers restrict immigration to appeal to these forces.
"New" new trade theory (NNTT) should have a similar effect. In NNTT, states engage in intra-industry trade, usually with states with similar endowments. Trade in automobiles between the U.S., Germany, and Japan is an example of intra-industry trade.
Trade occurs because consumers like variety, wanting the ability to consume Fords, Volkswagens, or Hon-das. Because trade is costly-firms have to find importers for their products, comply with different regulations, and pay to ship their product-only the most productive firms can afford to engage in trade. 20 Trade openness decreases the cost of trade, 12 allowing firms with lower levels of productivity to export, subjecting all firms-foreign firms, domestic exporters, and domestic-only producers alike-to increased competition in the domestic market. The least productive firms, which prior to the change in trade were producing for only the domestic market, lose market share to the new foreign competition and close. After trade liberalization, there is a higher average level of productivity than before. 21 As productivity increases, firms need fewer workers to produce as much and they often need more skilled workers, who tend to be more productive. 22 The closure of some firms in the low-skill tradeable sector, e.g. manufacturing and agriculture, also affects the remaining firms in those sectors and the low-skill nontradeable sector, e.g. construction and services. When firms in the tradeable sector close, they lay off both their native and their immigrant workers, who can now be employed in the remaining firms. The remaining firms no longer need to worry about the availability of lowskill immigrant labor (even if they did not rely upon immigrants in the past) because the layoffs have led to greater availability of native workers and lower wages. The remaining firms might still like additional immigrant labor but it becomes a lower priority for them and so they are less likely to support low-skill immigration. This dynamic helps explain why the service industry has not sought as much immigration as we might think they would have even though it relies on much immigrant labor. Policymakers can restrict immigration, then, to make others in the polity, including native labor and those who oppose immigration, better off without making non-tradeable firms worse off.
The converse also holds: greater trade restrictions should lead to more support for low-skill immigration. With the increase in restrictions, more low-skill and low-productivity 21 New firms that enter the market will be more productive, since only those more productive firms can compete with foreign firms. For a formal discussion, see Elhanan Helpman, "Trade, FDI, and the Organization of Firms," Journal of Economic Literature 44, no. 3 (2006): 596. 22 See Helpman, Itskhoki, and Redding above, n 18.
pro-duction happens in the domestic market. Additional firms will enter the low-skillintensive sector and/or existing firms will expand to fill the demand that was once covered by imports. As these sectors expand, wages for low-skill workers increase, leading to an increase in wages across the economy (assuming that labor can move relatively easily between sectors). The increase in wages will harm firms' profits and may even erase the gains that

B Firm mobility
The ability of firms to move production overseas, what I term "firm mobility," also acts as a substitute for low-skill immigration. Unlike trade, though, policy-makers have much less control over many aspects of firm mobility. Importantly, it must be technically feasible for the firm to move production to another location. Additionally, other states must be willing to allow foreign direct investment. investment treaty with another country, which would give firms greater legal protections in the face of potential expropriation, smoothing the path for home-country firms to invest in overseas production. With firm mobility, firms can either bring low-skill workers to their factories in wealthy countries or take their factories to low-skill workers in developing countries.
Which option they choose will depend on the relative ease of moving the factory. If the industry is immobile, firms will support open immigration. For example, much construction has to be done on-site. In contrast, manufacturing textiles or sewing t-shirts can be done anywhere in the world. Once firms move, they no longer support immigration at home, since they do not need the workers. The firms that do remain at home -whether because they are non-mobile like much of construction; they cannot afford to move; or are profitable without moving -can take advantage of the labor laid off when other firms move just like in the case of trade. As it becomes easier for more and more firms to move overseas, business support for low-skill immigration at home declines and policymakers can restrict immigration. In contrast, if it becomes more difficult for firms to move overseas, they will be more likely to support immigration and immigration policy should open.
The effect of firm mobility has become increasingly important with global supply chains. Most firms now rely upon production of at least some parts abroad. When firms out-source production abroad-either internally through a subsidiary like the Ford Motor Company's plants in Mexico or externally through a subcontractor like Apple's relationship with Foxconn-they need less labor at home. This will likely decrease their demands for immigration into the future.

C Productivity & technological change
15 Finally, technological changes that increase productivity will affect preferences over low-skill immigration. With the industrial revolution, technology has increasingly let businesses replace human labor with machines. Greater globalization has likely also played a role in speeding up the creation and adoption of labor-saving technology, as firms have sought to stay ahead of their competition. Increasing technological developments have also led to less support for low-skill immigration. Some firms simply need less labor and the labor they do hire needs to have more skills, 24 while other firms can take advantage of the excess labor that gets laid-off when another firm or sector mechanizes production to a greater extent. As productivity has increased, support for immigration should fall and with it, immigration policy should become more restricted.

D Integration of world markets and low-skill immigration
Altogether, then, greater integration of the world's goods and capital markets has led to greater disintegration of the worlds labor markets. Increased trade has meant that many firms that once supported low-skill immigration have closed their doors. Others have decided that if they can't beat the foreign competition, they will join them and moved their operations overseas, taking support for immigration with them. Finally, still others have mechanized production or adopted other productivity enhancements that mean they can do more with fewer workers. All three of these strategies have led to less support for immigration.
There are industries left in advanced industrial economies that still want low-skill immigration. These industries, like labor-intensive agriculture, construction, and hospitality, though have a hard time getting the immigration policy that they want because they no 24 See Helpman, Itskhoki, and Redding above, n 18. 16 longer have as many allies in other industries. Additionally, because there are fewer industries that need immigrant workers, a restrictionist politician can more easily find other policies that can appease these industries instead of immigration, such as tax changes, labor policies, environmental policies, and the like. Regardless, because much of the firm demand for low-skill immigration has declined with greater globalization, policymakers have been able to restrict immigration to appease to other constituents.

Post-War Era
As World War II was coming to the end, American and British policymakers began making plans for the post-war order. The goal of these plans was to create a lasting peace while preventing Communism from spreading further. The architects of these plans, Roosevelt was concerned about the impact refugees, "surplus" population in Europe and Asia, and "geopolitical problem children," minorities who were traditionally exploited for political gain by the great powers, would have on the stability of the world system. 28  While the ICEM was curtailing its mission, trade and capital mobility increased.
Throughout the 1950s and 1960s, the GATT had increased its mission, increasing its scope of tariff cuts and the number of participants. By the 1970s and 1980s, capital controls were increasingly removed. 41 The decline in trade restrictions led to deindustrialization throughout the advanced industrialized economies, as low-skill intensive industries moved to newly developing states, leading to the loss of manufacturing firms in those states. Among the remaining manufacturing firms, some moved production overseas either through foreign direct investment or subcontracting. Others increased the use of technology to replace labor. The remaining firms benefited from the loss of jobs in the manufacturing sector, employing the now laid-off labor without raising wages much: in the US, real wages have barely increased since the mid-1960s and throughout the OECD as a whole, median wages have only increased about 10% in the last 20 years. 42 Together, these developments shrank business support for immigration and led to 20 increased immigration restrictions. Figure 1 shows the relationship between trade and immigration policy in the "Western Core" of the liberal international order of Western Increased restrictions on immigration has likely dampened immigrant flows. During the same time as these restrictions have been put in place, the ability to migrate has increased as developing states have increasingly removed barriers to emigration 47 and air transportation has gotten cheaper and easier, leading us to expect that migration should have increased. Nonetheless, migration has remained stable as a percentage of the world's population; migration (including the number of undocumented migrants) is about 3% of the world's population today, the same that it was in 1960. 48 We can see the effect of these restrictions when we examine survey data about the desire to move; about three times as many people would like to migrate as actually migrate. 49 Of those that do migrate, fewer are going to the Western Core today. Instead, about half of migrants today migrate within the developing world or to wealthy autocracies. 50 Even undocumented immigration is a sign of the increasing restrictions, as it shows that many more people would like to enter a country than legally can. 51 Thus, in a world without as many restrictions, we would have expected immigration flows to be larger; indeed, Peters finds that these restrictions have a significant 23 effect on migrant flows. 52 This relationship between trade and immigration became stronger when capital restrictions were lifted. Figure 2 shows the relationship between trade and immigration during the Bretton Woods Era, when capital was more restricted, and after the Bretton Woods period when capital restrictions were largely removed; the dots represent each country-year in the data and the lines are the line of best fit for the data. At the same time that wealthy nations removed restrictions on capital moving out of their economies, developing nations were increasingly allowing foreign firms to engage in foreign direct investment. 53 The combination of tariff reductions and decreasing restrictions on capital mobility and foreign direct investment made it more attractive for firms to move production. Now they could produce overseas, import their products back to the home country, and move their money from one production site to another. As more firms moved some or all of their production overseas, their need for low-skill labor decreased and with it their support for immigration.
Policymakers, then, increased immigration restrictions to appease other constituents. 54 Within the EU, we have seen a similar effect of the creation of the integrated 52 See Peters above n 3, at 293. 53 Sonal S. Pandya, Trading spaces (New York, NY: Cambridge University Press, 2014). 54 As Peters, above n 3, shows, the relationship between immigration, trade, capital, and FDI policies is robust to the inclusion of additional variables, many of which fail to reach conventional levels of significance.

24
European goods and capital markets on firms support for immigration and immigration policy; how-ever, states have given up their control over immigration from within the EU. As producers in any one country of the EU can no longer get trade protection from any other, they were subject to greater competition from low-wage areas, such as Southern and then Eastern Europe. However, because these firms can easily move within the union, some firms offshored pro-duction to these states. Finally, those firms that were immobile-primarily agriculture and services-have been able to rely on intra-EU migrants. The effect, then, of the integration of the EU has been "migrant diversion," similar to trade diversion, in which opportunities for immigration from outside the EU has declined as opportunities for immigration within the EU have increased.
Thus, as globalization and regional integration has increased, support for low-skill immigration from businesses has decreased. Overall, this has led to increased restrictions on low-skill immigration. As I show in the next section, it also opened up space for political entrepreneurs on the right to pursue hardline immigration restrictions as a political strategy.

Case of the Republicans
One question that occurs when we think about the rise of immigration restrictions, especially in the U.S., is how have the nativist wings of conservative parties taken over immigration policy? Historically, business interests and nativist groups have coalesced in political parties on the right. For example, in the U.S., the Republican party gained much of its support in the 1850s from smaller, parties and groups, like the Know-Nothing Party, that were anti-immigration. Yet, not long after forming the Republican Party quickly embraced immigration to appease Northern business interests. In the late twentieth and 25 twenty-first centuries, we have seen a reversal. Now, it appears that the Republicans have abandoned business on immigration, but not, crucially, on most other issues. I argue that this movement has less to do with the politicians abandoning the position of business on immigrant than the fact that business has stopped supporting immigration to the same extent it once did.
To  Democrats and Republicans converged on immigration policy to some extent in the middle of the twentieth century. In part, this was due to the limited reopening of immigration during and after World War II, when both parties realized that the pre-war status quo was too restrictive. By the 1970s, though, the parties diverge again, but this time with the Democratic Party as more supportive of immigration.
To understand this divergence, I examine how support by each party changed based on the change in the national tariff rate and the ability to move production abroad from 1970-2000. 61 The national tariff rate is the same tariff rate used above. The ability to move production abroad is based on how many English-speaking countries in the world freely allow FDI to enter their country. 62 I examine the English-speaking world as it is easier for firms to enter an economy in which the workers and customers speak English.
The ability to move into other countries is used because the U.S. had almost no restrictions on outward foreign direct investment during this time. To examine the relationship between voting, trade openness, and offshoring, I regress each senator's average vote for the year on the tariff rate, the ability to offshore production, and several controls. 6363 Figure 4 shows how the predicted support for immigration by Democrats and Republicans change as trade openness (left panel) and the ability to offshore production (right panel) increase, respectively. We find that trade and the ability to offshore production negatively affects both parties support for immigration. Republicans over this time period are actually slightly more likely to support immigration when trade was more restricted and the ability to offshore was low. But as these impacts increase, we see that this reverses and 60 Ibid. 61 Time range is due to data availability. 62 Data are from Pandya, above n 34. 63 For more details, see the Appendix, table 1.

28
Republicans become less likely to support immigration than Democrats. Republican support for immigration is also affected by changes in productivity. As firms increase their productivity they need fewer worker and so are less likely to support open immigration. The effect of increasing productivity is especially strong on senators who represent the agricultural sector. While agriculture in general is more supportive of immigration, as it has become increasingly mechanized, it has needed fewer workers and decreased demands for immigration. Thus, for Republicans who represent agriculture that needs fewer workers-crops like wheat, soybeans, and cotton-their support for immigration is much lower than for Republicans who represent agriculture like fruits and vegetables that still rely on a large workforce. 64 We see that as business support for immigration has declined, so has Republican support for immigration.

Figure 4 about here
The declining support for immigration by firms also helps explain some of the changes The fact, then, that businesses have stopped supporting low-skill immigration at the same rate as they did in earlier eras means that politicians can cater to other constituents on this issue. In practice, this has meant that the Republican Party in the U.S.
increasingly has catered to nativists on this issue. Whereas Freeman once argued that politicians were hamstrung in how they discussed immigration by the strictures of political correctness, 68 this is clearly no longer the case, at least in the U.S. It is likely that as Republicans increasingly catered to nativists, they push the boundaries of acceptable conversation, allowing room for a politician like Donald Trump to be part of acceptable politics. By providing this opening to politicians who were on the far-right of immigration policy, center-right politicians likely opened the door for the far-right in the U.S., Europe and elsewhere.

IV Conclusion
The Liberal International Order (LIO) that the U.S. and U.K. constructed after World 30 War II has long has its critics but until recently, those critics were not the leaders of the founding countries. The LIO has come under attack at a time when its benefits have accrued to the well-off in the developed world at the expense of the middle and lower classes. 69 Yet, one of the main villains of the backlash to globalization and the broader LIO has been immigration, which was never been part of the LIO.
I argue that it was, in fact, the sequencing of the LIO that led to increased restrictions on immigration. Increased trade led to the closure of firms that intensively used immigrant labor, as they were no longer competitive. The ability to move production overseas meant that firms could now bring their capital to labor instead of fighting to bring labor to capital. Finally, the resulting technological developments have meant that employers can do more with fewer workers. Together this has sapped the support of businesses for lowskill immigration. As businesses pulled back their support, anti-immigrant forces have become ascendant. Data on trade barriers, capital mobility, and immigration policy across the major economies of the Western core of the LIO demonstrate how increasing trade and firm mobility have led to increased restrictions on low-skill immigration. Further, data on voting by U.S. senators shows that as trade and firm mobility increase, Republicans, the right party in the U.S., have moved further in the anti-immigrant direction as business support for immigration has decreased.
This sequencing of the LIO, with trade and capital opened first, was meant to protect native labor from competition. Paradoxically, it likely made competition worse. One of the major points of contention with globalization, primarily on the left, has been that businesses in the developing world gain an unfair advantage because they do not have to conform to the same labor and environmental standards that businesses in the developed world do. Yet, the playing field would have been leveled, to some extent, if more immigrants had been brought to the developed world, instead of allowing production to move to the developing world. Businesses using immigrant labor in the developed world would have been required to conform to labor and environmental standards. Further, we know now that immigrants are largely complements for native workers, taking on the more routine tasks while allowing native workers to put their skills to use and all the while cutting total labor costs. 70 Greater immigration, thus, could have helped protect the manufacturing sector in the developed world, by making it more competitive, while not harming native workers. More high-quality contact between immigrants and natives could have increased understanding, leading to less nativism 71 and perhaps, less populism. Additionally, with greater understanding, natives would have been better able to organize immigrant workers in the developed world, further increasing labor standards, and even spreading unionism back to developing countries through social remittances. 72 It is then possible that the absence of immigration in the LIO helped sow the seeds of its demise rather than protecting it. 70 See Peri and Sparber above, n 7. 71 See Allport above, n 6. 72 Margaret E Peters and Michael K Miller, "Emigration and Political Contestation," Unpublished Manuscript. UCLA, 2020.