Margaret E Peters. Forum. Volume 17, Issue 4. December 2019.
Introduction
When President Trump came into office, he promised he would make America great again for the American worker. He touted that he would protect worker by renegotiating our preferential trade agreements and restricting immigration. Thus far, he has claimed to make good on both of these promises.
Yet, as I argue in my book (2017), Trading Barriers: Immigration and the Remaking of Globalization, Trump can’t have both trade protection and immigration restrictions because businesses will fight against it. Trade protection should increase the production of goods in the US. With tariffs, Trump’s favorite form of trade protection, prices of foreign goods become more expensive. As foreign goods become more expensive, Americans consume more domestically made goods, raising demand. The increase in demand, in turn, should lead to expanded production—whether from existing firms or new entrants—and with that more jobs. Increased demand for labor should eventually reverberate across the economy and raise wages and labor costs for all employers. These employers, then, push for increased immigration to bring labor costs back down. Because employers are relatively powerful—especially in this administration—we expect immigration to open.
As I show in my book, this relationship has held, more or less, throughout US history and in much of the developed world. Take, for example, 19th century America. During that time, the US consistently had relatively high tariffs. The goal of the tariffs were both to generate revenue and to protect developing industries. The tariffs worked; many new industrial firms were created or expanded. As these firms expanded, so too did wages, leading to demands for immigrant workers. The federal government did little to restrict immigration until late in the 19th century; even then most European immigrants who wanted to come to the US could.
After World War II, the US increasingly opened its borders to foreign goods through reciprocal trade agreements through the General Agreements on Tariffs and Trade (GATT), the precursor to the World Trade Organization (WTO). With this openness to trade, many manufacturers closed their doors. With their closure, millions were laid off and wages stagnated. Employers no longer demanded as many workers and support for immigration fell. It wasn’t the case that employers did not want cheap labor; they already had it and saw little need to spend political capital fighting for more.
President Trump came into office saying that he would reverse this trend. He has placed tariffs on many goods from several countries, most notably on China. He has renegotiated NAFTA, creating the USMCA. At the same time, he has used executive action to try to restrict immigration; most notably he engaged in the morally reprehensible family separation policy and now the “Remain in Mexico” policies that have led to a decline in asylum seekers. He has also dramatically cut the number of refugees resettled in the US.
Does this mean that the Trump administration has found a way out of the relationship between trade and immigration that has affected other leaders? I offer that there are several hypotheses why businesses have not demanded more immigration, even in the face of increased trade restrictions. The first hypothesis argues that trade restrictions have not led to much job growth. In contrast to the 19th century or even the 1950s, American firms rely increasingly on exports. Because other countries have retaliated with their own tariffs, the US may have lost more manufacturing jobs with the tariffs than have been created. A second hypothesis is that the labor market is not actually as tight as the unemployment numbers make it seem. The labor force participation among prime age adults (25-54) is still not where it was in 2001. This slack in labor force participation could be the reason why both wage growth has still been slow and employers have had not had too difficult a time finding employees. A third hypothesis is the rise of automation means that firms simply don’t need the labor that they once did. Finally, it’s possible that for all his talk, Trump actually has not restricted immigration much.
Through the rest of this essay, I will further elaborate the theory of my book. I will then discuss how the theory has applied over the history of the US. Finally, I discuss my four hypotheses for why even President Trump has not (yet?) changed this relationship.
Trade, Firm Mobility, Automation, and Immigration in Theory
I argue that trade openness, firm mobility, the ability to move production abroad, and productivity increases, and automation, decrease political support for lowskill immigration in wealthy countries. Wealthy countries are the places that most immigrants want to move to—they must decide whether and how to regulate immigration.
When crafting immigration policy, I assume that policymakers, like the President, Member of Congress, and the like, do so with an eye to ensuring their reelection. To stay in office, policymakers must have both elite and mass support, even in a democracy like the US. Politicians in the US need some combination of support from business groups, like the Chamber of Congress, other industry groups, and corporations; organized labor; other special interest groups; and voters. Special interest groups can provide knowledge about the policy area (Grossman and Helpman 2001; Hall and Deardorff 2006), campaign contributions (Grossman and Helpman 1994), and can even help get out the vote (Hertel-Fernandez 2017). Politicians may have their own position on an issue or care about party cohesion and so will balance their constituents and special interests against their and their parties’ beliefs as well.
Special interest groups and voters’ positions on immigration run the gamut from pro-immigration to anti-immigration. Pro-immigration groups often include prior immigrants (Goldin 1994; Freeman 1995) and human rights and other humanitarian organizations (Freeman 1995). Historically, they also included business interests, but as I will argue these interests have declined in number over time. The anti-immigrant side has consisted of (at times) native labor (Briggs 2001; Fine and Tichenor 2009), tax payers worried about the fiscal costs of immigrants (Neuman 1993; Hanson, Scheve, and Slaughter 2007), and those who oppose immigration due to the cultural changes that immigration brings (Higham 1963; Cohn 2000; Hainmueller and Hiscox 2010). Voters are thought to approve of or oppose immigration based on their membership in these groups.
Historically firms have been the major supporter of immigration. While prior immigrants may greatly support immigration, they may not yet be citizens, and thus not voters, and even once they obtain citizenship, they typically vote at lower-rates than the native-born. Human rights and humanitarian groups typically do not have great resources either. Without firms, anti-immigrant groups will likely have more strength than pro-immigrant groups.
Firms are strategic in their using their political capital, or what I will refer to from here on out as lobbying. Lobbying is not free—it costs money to hire lobbyists, whether in house or at a firm—and political contributions and get out the vote drives take money. This is money that could be spent on other activities in the firm—like adopting new technology—that would make the firm more competitive or on the firm owners themselves. It could also be spent lobbying for other issues than immigration that might be easier to achieve or save the firm more money, like tax policy. Firms, then, are not going to supply capital in excess of the expected benefit they will get from an immigration policy change.
How much a firm will benefit from a change in immigration depends on the type of immigrants that the policy affects. A manufacturing firm, for example will not lobby for a policy change that only leads to more agricultural workers and vice versa. In this paper, I’m primarily concerned with low-skill immigrants—those who would work in jobs that need relatively few formal skills. High-skill intensive firms, like technology firms, are unlikely to lobby much for low-skill immigration; instead they are interested in high-skill migration. Similarly, farmers are unlikely to lobby for high-skill immigration.
Firms will have different preferences over low-skill immigration depending what other options they have to produce besides using low-skill labor in their home country. One option is the adoption of technology. The more firms can replace humans with machines, the less they are likely to support low-skill immigration. A second option is to move production overseas to a developing country with lower labor costs. If firms are no longer producing in the home country, they no longer have need of immigrant labor there. Trade will affect these preferences by changing the number of firms that rely on low-skill labor. In developed countries like the US trade protections should increase the production of low-skill labor intensive goods. Trade openness has the opposite effect: less production here in the US. Firms respond to increased trade openness by automating to use less labor, moving production overseas, or closing. Regardless, the number of firms that support low-skill immigration should decline. The decline in business support, then, gives anti-immigration forces relatively more influence over immigration policy and leads to more immigration restrictions.
Trade’s Effects on Business
In classic Ricardian trade theory, states trade based on their comparative advantage. Wealthy states are abundant in capital and high-skill labor and goods that use much of those factors of production in their production (high-skill intensive goods) are relatively cheap while those that use the scarce factor of production, low-skill labor (low-skill intensive goods) are relatively expensive. Developing countries are opposite: abundant in low-skill labor but scarce capital and high-skill labor.
Protecting their markets from trade means protection for the production of goods that use the scarce factor. For developed countries this means low-skill intensive goods. When developed states erect trade barriers, the low-skill intensive sector should expand, as goods from overseas become expensive. As this sector expands, it pulls in labor from the rest of the economy. As long as labor is relatively mobile, this means that wages should increase across the economy. Now, even firms that do not engage in trade—like services—see their wage bills go up.
Immigration can help firms lower their wage costs, either directly or indirectly. For jobs that are routine or only require manual labor, an immigrant, even one who does not speak the language, can easily replace more expensive native labor. For other jobs, it is difficult for immigrants to replace natives because the job requires language skills or other customer facing skills requiring cultural knowledge (Peri and Sparber 2009). All businesses have a mix of routine/manual work and less routine and cultural specific work. Immigrants can do the more routine work freeing expensive labor to do the less routine work, lowering the total wage bill even if they do not replace natives (Peri and Sparber 2009).
With trade protection, then, firms from across the economy should lobby for increased low-skill immigration to lower their wage costs. Because firms are relatively powerful, we expect that they will get at least some of their demands and that low-skill immigration should open.
When wealthy states remove trade barriers, we expect to see the opposite happen. As trade opens, low-skill intensive products from overseas become cheaper and, thus, consumers will buy more of them. This will decrease the demand for domestic goods, and with it, their price. Firms can either cut costs to remain in business, usually by cutting labor costs and/or automating production, or close. Those that close can no longer lobby on immigration. Those that lay off labor and/or automate no longer need as many workers. Either way, the demand for low-skill labor declines and with it the demand for low-skill immigration.
“New” new trade theory (NNTT) has a similar effect. In NNTT, states engage in trade within the same industry, or intra-industry trade, usually with states with similar endowments. The automobile industry is a classic example of intra-industry trade: the US ships Fords, GMs, and Chryslers abroad while also importing Volkswagens, Hondas, and Kias. States engage in this trade because consumers like variety. Yet not all firms can engage in trade. Trade has both fixed costs—like finding an importer and complying with regulations—and variable costs—like the costs to ship a good. Because of these costs, only the most productive firms, those who have the most money left over after making the good, can engage in trade (Melitz 2003). Trade openness in the domestic market decreases the cost of trade, by lowering tariffs, reducing regulatory barriers, or reducing the cost of shipping. This allows firms with lower levels of productivity to export to the domestic market. Yet, by allowing more firms to export into the domestic market, trade subjects all firms, the preexisting foreign import firms, domestic firms that export and produce for the domestic market, and domestic-only producers alike, to increased competition. The least productive firms, which prior to the change in trade protection 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 (Helpman 2006). As productivity increases, firms need fewer workers to produce as much and they often need more skilled workers, who tend to be more productive (Helpman, Itskhoki, and Redding 2009).
When firms that used a lot of labor close, they lay off their workers. These workers, both native-born and immigrants alike, can then be employed in the firms that still remain in the economy. Now these firms have less need for lowskill immigrants, too, and decrease their lobbying for immigration as well. Policymakers can restrict immigration, then, to make others in the polity, including native labor and those who oppose immigration, better off without making the remaining firms worse off.
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, politicians have much less control over firm mobility. To move production, it first must be technologically feasible to do so. This means that some activities are simply not offshored. These include activities that take advantage of natural resources in one location—like oil or minerals or particular growing conditions—and activities that must be done near the customer—like concrete or construction. Further, it must be possible to control the firms’ agents overseas. Prior to the development of the telegraph, and later telephone and internet, it was difficult to have timely communication with agents in another location, making production abroad difficult. Finally, firms have to be allowed to invest in other countries, something that was more of a concern in the past (Pandya 2013).
There are some factors that policymakers can control. They can reduce their currency’s convertibility, which makes moving money from the parent company to the subsidiary more difficult. Alternatively, they could sign a bilateral investment treaty which should make overseas investment less risky and, therefore, cheaper.
With firm mobility, firms can take their production to the workers rather than bring the workers to the production site. Once firms move, they no longer support immigration at home, since they do not need the workers. Non-mobile industries can benefit from native laid off workers when mobile firms move in the same way that firms benefit when other firms close due to increased foreign trade. As it becomes easier for more and more firms to move overseas, business support for low-skill immigration at home declines and politicians can once again restrict immigration to make other groups happier. 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.
Automation and Productivity
Since the industrial revolution, technology has increasingly allowed businesses to exchange expensive human labor for relatively cheap machines. Automation, on its own, should decrease the demand for low-skill immigration as fewer workers are needed. When firms automate, they typically need more skilled labor as well, who can better understand the machines that they work with, further decreasing the need for low-skill labor. It is also the case that globalization has led to increased adoption of labor-saving technology. The demand to compete with firms with lower labor costs likely led some firms to develop and adopt new technology. Regardless of where the technology came from or why it was adopted, when firms adopt laborsaving technology, they no longer need as much labor. The labor that they lay-off can work in areas which have not yet automated. Together, then, this also lowers firm demand for low-skill immigration. Policymakers can again restrict immigration to make other groups in the polity happier without making firms worse off.
Globalization and Low-Skill Immigration
Greater globalization over the last two centuries has led to less demand for low-skill labor by firms in wealthy countries. Some firms that used to lobby for lowskill immigration have closed. Others have moved overseas. Still others adopted labor saving technologies. There are still industries left in wealthy countries like the US which want low-skill immigration, like labor-intensive agriculture and construction. Yet, this decline in total demand for low-skill immigration has allowed politicians to increasingly restrict immigration to make other groups happier.
Globalization and Immigration Policymaking in the US
The US is a nation of immigrants, it is true, but we have not always been a country of open immigration. To understand how immigration policy has changed, I coded US immigration policy, along with the immigration policy of 18 other countries. I coded policies on which immigrants could enter the country, what rights those immigrants have, and enforcement policies and then combined those dimensions into a single index.3 Figure 1 shows how US policy has varied over time along with the total number of legal permanent residents divided by GDP
Figure 1 shows that in the 19th century, the US was very open to immigration. In the early 20th century, the US closed its doors to immigrants first with the Literacy Act of 1917 and then with the Quota Acts of 1921 and 1924. Immigration remained highly restricted until 1965, when it was opened somewhat with the Immigration and Nationality Act of that year. Since then, immigration has been increasingly restricted.
Some argue, however, that immigration policy must have been opened to a much greater degree with the 1965 Act given our increasing percentage of foreign born.4 Yet, this argument ignores that native birthrates have been declining. Additionally, we have to ask what the proper counterfactual measure of business demand is. If businesses demanded labor at the same rate as they had in the past and native born birth rates still declined, we should have much greater numbers of immigrants per dollar of GDP than we did in the past. Instead we have many fewer, even as our economy has grown.
Figure 2 shows how US immigration and trade policy—as measured by 1 minus the tariff rate—have varied across history. The eyeball test suggests that the US follows my theoretical argument. While they bounced around depending on which party was in power, tariffs were much higher in the 19th century as the federal government sought to develop manufacturing and needed tariffs for revenue. At the same time, immigration as noted was almost completely open for much of the 19th century. Starting in the early 20th century and continuing in the post-World War II era, tariffs began to come down and are now about 2%, with many goods coming in duty free.5 During this same time, immigration was increasingly restricted.
The Americanization of the US Economy
Increasing openness to the rest of the world wasn’t the only factor that affected US immigration policy in the 19th century. The “Americanization” of the economy also played a role. In the early days of the Republic, the US was a series of regional economies, rather than one integrated economy. One of the major factors to change this was the ability to transport goods cheaply among the different regions of the US. While turnpikes and canals were important improvements, the most important transportation advancement of the nineteenth century was the railroad. Railroads in the US grew from nothing in 1830 to over 400,000 miles of rail by 1920 (Kim 1995, p. 885).
Once the railroad connected an area to the rail network, it had much greater market access. Increase rail access was a blessing for some producers but a curse for others, as it led to price convergence (Kim 1995, p. 886). Small firms, which had a local monopoly, were often pushed out of the market once railroads were expanded into their region (Engerman and Sokoloff 2000).
The most striking example of the effect of the railroad on businesses and their support for immigration was the completion of the transcontinental railroad in 1869, linking the industrial East and Midwest to the West. The completion of the railroad had two major effects on the economy of California. First, it led to massive lay-offs of both Chinese and White workers, pushing down wages (Saxton 1971; Boswell 1986). Second, and perhaps more importantly, it led to a recession as California producers closed when they had to compete with Eastern producers. Eastern producers had much lower labor costs; Chinese workers in California were paid only slightly less than White workers in the East and White workers in California often made twice as much as those in the East (Boswell 1986, p. 363). I argue that after the completion of the railroad there were fewer firms to push for Chinese immigration and for the remaining firms, there was less need for new immigration as the firms could use the laid-off White and Chinese labor.
To examine this hypothesis, I regressed how senators voted on immigration bills on the percent of Chinese immigrants in the state, an indicator denoting the time period, and the interaction between the percent of Chinese immigrants and the indicator. As Chinese immigrants were welcomed (or at least not prohibited) at first, this analysis examines when the Chinese population in a state moved from being a benefit to a political liability. I broke the time period as pre and post the completion of the Transcontinental Railroad (as either 1869 or 1870, since the railroad did not reach Oakland until November 1869) and several placebo periods, including a few years before and after the completion of the railroad; 1881, the year before the Chinese Exclusion Act was first passed; 1897, a randomly chosen year; and 1914, the start of World War I.
Figure 3 shows the coefficient on the percent of Chinese immigrants in each state and the interaction term. Prior to the completion of the rail (for breaks at both 1869 and 1870), the effect of Chinese immigration is positive and statistically significant; states with more Chinese immigration—states in the West—were much more supportive of immigration in general, including on bills on immigration from China. After 1869 or 1870, the percent of Chinese has no effect or a slight negative effect on voting behavior. The same difference is not found if we split the time period before the completion of the railroad or long after it.
Thus, something happened in 1869 or 1870 that changed the way senators in the West voted on immigration. Prior to the completion of the railroad, Western firms were protected by high transportation costs. Production expanded, leading to an increased need for labor and increased demands for immigration. After the completion of the railroad, firms lost this protection and faced increase competition, leading to the closure of some firms. Remaining firms no longer needed more immigrant labor because labor costs were down due to the lay-off by the railroads and the closure of other firms and likely chose to lobby on other issues or to not lobby at all, allowing anti-Chinese immigration legislation to be passed.
In addition to the changes in transportation, intra-US firm mobility increased in the second half of the 19th century as well. The telegraph and later the telephone allowed for quick transmission of information between the branches of a firm. Typewriter allowed for greater written communication, giving managers greater control and information (Yates 1993). These technological and managerial changes allowed firms to produce across the US.
Changes in the US banking and financial sector also affected firm mobility. Prior to the Civil War most firms had to self-finance capital improvements (Engerman and Sokoloff 2000). After the Civil War, changes in the structure of the banking sector as well as the development of the commercial paper market and stock markets allowed firms to expand both in size and geographically (Davis 1965; Engerman and Sokoloff 2000). Changes in banking laws also allowed capital to flow more freely from the Northeast to areas scarce in capital (Davis 1965; James 1976). All in all, it became easier for businesses to move across the country.
Finally, the industrial revolution itself played a role. There were huge increases in productivity throughout the economy. One area where this was particularly pronounced was agriculture. Mechanization of agriculture decreased the work need to produce an acre of corn by a factor of 5 between 1855 and 1894 and for wheat by a factor of 25 between 1829 and 1894. While not all agriculture was mechanized, mechanization released large amounts of labor. Farmers as a percent of US population declined from a high of 16 percent in 1870 to less than 10 percent in 1925 and less than 6 percent in 1950.8 Farmers were less likely to hire labor and farmers’ children were more likely to leave the farm for the factory, reducing demand immigrant labor across the economy.
Technological development also affected manufacturing. The first shift in industry from artisanal shops to factories and assembly lines in the 1830s to the 1880s led to an increased need for relatively unskilled labor, which likely increased support for immigration. The next shift from factories to continuous and batch processing methods beginning in the 1890s led to an increased need for skilled labor at the expense of unskilled labor (Goldin and Katz 1998).
Together, then, these changes would have decreased the need for immigrant labor. Firms that used to be shielded from competition faced increased trade pressures from areas of the US with cheaper labor forces. Other firms moved to those areas, especially to the South, which had much lower labor costs (Wright 1979). Finally, still others increased their productivity through mechanization. Firms support for immigration declined, which allowed anti-immigrant groups at the time—organized labor in the AFL and nativists—more power over immigration policy, leading to restrictions.
The Post-World War II Globalization
With the passage of the Reciprocal Trade Agreements Act (RTAA) in 1934, Congress began opening US trade policy. The RTAA and its successor, Trade Promotion Authority or Fast Track, allow the president to negotiate reciprocal trade agreements with other countries that lower tariffs, and after 1974 non-tariff barriers (NTBs) as well, that Congress can only vote up or down. This form of trade openness was effective: leading to the creation of the GATT and later, the WTO as well as a myriad of preferential trade agreements like NAFTA.
After World War II, there was an increase in firm mobility as well. For most US firms the obstacle for successful off-shoring was not the US government, but foreign governments. The US has had few controls on capital. Instead, foreign governments had to be open to FDI, which developing countries increasingly were starting in the 1970s (Pandya 2013). As we have seen, increasingly firms have become multinational firms, with production around the globe. Other firms have increasingly outsourced some production to firms overseas as well.
The immediate post-World War II politics suggest that trade and capital policy were set first, which then had path dependent effects on immigration policy. Even before the war ended, US policymakers were developing plans for the peace, including opening trade, through what eventually became the GATT/ WTO, and recreating the liberal international order in finance through the IMF. These policies, along with the Marshall Plan and America’s liberal capital policies which allowed firms to invest overseas, were implemented in hopes of rebuilding Europe and Japan; preventing Communism from spreading further; and binding European countries together to prevent another war (Hull 1948; Irwin 1998; Ikenberry 2001; Barton et al. 2006).
On immigration, however, the focus was on how to maintain the status quo while dealing with refugees created by the war. Yet instead of international cooperation on economic immigration, similar to the cooperation on trade and financial flows, the US and other Western nations states created the Intergovernmental Committee for European Migration, which helped move migrants from Europe to the New World, but not to the US. Thus, at the end of World War II, the US decided to open trade and allow firms to move, while maintaining its immigration policy
In addition to changes in the global economy, there have again been major increases in productivity in both agriculture—including the “Green Revolution”—and industry. One example of the effect of mechanization on agriculture’s support for immigration was the mechanical cotton harvester. As late as World War II, it still took about 100 man-hours of work per acre of cotton (Census Bureau 1975). Even with the slow adoption of the cotton picker in the South, the number of man-hours need per acre of cotton fell by half by the early 1960s and by 1970, it had fallen by half again to 24 man-hours.
The mechanization of cotton may have led to the end of the Bracero program, the agricultural guest worker program between the US and Mexico in 1964. By 1964, about 85 percent of total US production of cotton had been mechanized, but the percent of cotton mechanized in western states was much higher: ranging from 90 percent in Texas to 98 percent in California and Arizona. Even cotton farmers in the South, who had not used the Bracero Program to a great degree, became more supportive of immigration restrictions with mechanization, since they no longer had to worry about Western farmers stealing their workers. With the loss of the cotton lobby, policymakers may have felt less pressure to maintain the Bracero Program and allowed it to end in the face of opposition from organized labor.
Industry too has seen large gains in productivity over the last 75 years. Increasingly, businesses have replaced human labor with machines across the US economy. This too has led to less support for low-skill immigration (although likely for more support for high-skill immigration) as companies no longer need as much labor.
Finally, these dynamics can help explain the rightward turn on immigration by the Republican Party. Figure 4 shows how support for immigration by the average Republican and average Democratic senator has changed. The data confirm the conventional wisdom about the parties support for immigration. Republicans opposed immigration to a greater degree when the party first came into existence in the 1850s. After the Civil War, Northeast producers came to dominate the party. They greatly favored open immigration to keep their labor costs down and with their support, the Republicans were able to keep immigration open with limited restrictions in the late nineteenth century. Democrats during this time period increasingly represented organized labor, which opposed immigration. Additionally, during the late nineteenth and early twentieth century, Southern Democrats moved from supporting immigration to opposing it: having realized that most immigrants would not move to the South, their racist ideology combined with their desire to maintain their low-wage advantage over the Northeast and Midwest led them to oppose immigration (Peters 2017).
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. Republicans have traditionally been the party of business, so when business support for low-skill immigration declined, so too did the support of the Republican party. The declining support for immigration by firms also helps explain some of the changes in Democrats support. Yet, the increase in their support seems to come from increase in the size of the foreign-born population. Democrats have long represented immigrants; thus, as the size of the immigrant population has increased, the Democrats have become more supportive of immigration. It also appears that the size of the welfare state matters: Democrats from states with larger welfare programs tend to be more supportive of immigration This is consistent with the embedded liberalism hypothesis: populations are more supportive of openness to the global economy, including immigrants, when the welfare state is there to protect them (Ruggie 1982).
Donald Trump, Trade, and Immigration
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 US increasingly has catered to nativists on this issue. Donald Trump in many ways represents the height of this trend. He ran on a virulently anti-immigrant campaign, yet also on increasing trade barriers. In office, he has enacted tariffs on many goods and made many executive orders on immigration. Can he have his cake and eat it too or will business interests demand more open immigration?
Hypothesis 1: Trade restrictions have not brought back that many jobs
For trade policy to have an effect on immigration policy, it must be the case that the trade policy is actually effective at changing what is produced within a country. So far, Trump’s tariffs have been relatively ineffective. Figure 5 shows the trade flows of exports and imports into the US on the left (Bureau of Economic Advisers 2019) and the change in manufacturing jobs on the right (Bureau of Labor Statistics 2019a). First, we see that since Trump first put tariffs into place in January 2018, both imports and exports have been relatively flat, suggesting the tariffs have had little effect on trade. But, perhaps these statistics hide import changes in what is traded and that these changes have had an effect on jobs. Yet, we see in the graph on the right that manufacturing jobs have also been relatively unchanged with the changes in trade policy. These data suggest at least some support for this hypothesis, namely that the changes in US trade policy have not (yet) had an effect on the market structure of the US economy that would change businesses’ support for immigration.
Hypothesis 2: Labor market isn’t as tight as we would think
For businesses to demand more immigration the labor market has to be tight enough that some employers cannot find the workers they need and for most others, wages have to be rising. While the unemployment rate in the US stands at 3.5% as of November 2019, it is possible that the labor market is not as tight as that headline rate suggests. Figure 6 shows the employment rate for prime age workers (25-54 years old; seasonally adjusted) on the left (Economic Research Division, Federal Reserve Bank of St. Louis 2019) and the median weekly wage (constant 1982-1984 dollars) on the right (Bureau of Labor Statistics 2019b). The percentage of prime age workers that are employed is close to what it was before the great recession started. Nonetheless, there are likely more people to bring back to the labor market; the prime age employment rate was about 2% higher in 2001. If the economy were to get back 81% prime age employment, that would equal about another 2 million workers today. This suggests that there might still be slack in the labor market and that businesses are still able to get the workers they need.
Wages tell a similar story. As can be seen on the right, median wage growth has been sluggish, with workers only seeing a few more dollars in their paychecks each week. This suggests that businesses are not facing much wage pressure. Together, the fact that there may be slack in the labor market and that there is little wage pressure means that businesses may not need additional immigration to stay profitable. Thus, they may not be willing to spend much political capital on immigration.
Hypothesis 3: Automation
Even if the tariffs were having an effect on trade flows and if the labor market was tight enough to see real changes in wages, it could still be the case that increases in trade protection would not have an effect on immigration if we saw an increase in automation. Businesses increasingly have the ability to automate functions that once had to be done by a human. For example, the steel industry produces more today than it did in 1970 but does it with less than a third of the workforce (Peters 2017, p. 106). Thus, it is possible that even if the tariffs did work to increase production of US made goods, that businesses would simple produce more with automation rather than hiring many more workers.
Hypothesis 4: Trump hasn’t really restricted immigration that much
My final hypothesis is that for all of Trump’s tough talk on immigration, he has not changed the inflows of immigrants much. Figure 7 shows the inflows of all immigrants, as well as by category, and the enforcement actions. Immigration data is for the fiscal year, which runs from the fourth quarter the year before through the first three quarters of the year. The data is currently only suggestive because the Department of Homeland Security has yet to publish data for fiscal year 2018 for many categories. Nonetheless, we can examine the actions of Trump’s first 9 months in office and in some cases his actions through the first year and 9 months.
Importantly for business demand for immigration, the number of temporary workers and their family members has been pretty flat H2A (agricultural workers) and H2B (nonagricultural workers), the two visa categories that represent most of the low-skill workers coming into the mainland US, have seen some fluctuation but are not much lower than they were under Obama. Similarly, the number of newly arrived legal permanent residents, those who move here as Green Card holders directly from overseas, has also remained more or less the same. Thus, there does not seem to be a major change in voluntary migrant flows to the US so far.
We have seen a much larger change in the number of forced migrants admitted to the US (Figure 7, middle panel). Most importantly, the numbers of resettled refugees—those people the US settles directly from abroad—has been dramatically reduced. Resettled refugee policy is one of the areas of immigration policy in which the President has more discretion, as he is allowed to set the ceiling. For FY2020, the Trump administration has once again reduced the ceiling to 18,000. Forced migrants can also come to the US directly and ask for refugee status, or what is known as affirmative asylum. At least as of FY 2018, grants of affirmative asylum had increased. This trend is unlikely to hold, however, as the Trump administration’s policies like “Remain in Mexico” will likely make asylum seeking more difficult. While these policies harm America’s international standing, they nonetheless have a relatively small effect on the total number of immigrants entering the US as refugees and asylum seekers are a relatively small part of the flow.
Finally, is it possible that Trump has increased enforcement, reducing the number of undocumented immigrants both by stopping them from entering the US and by deporting those already here? It does not appear that enforcement actions have greatly increased. At least through FY2017, apprehensions at the border and deportations (both removals and returns) were down significantly from the heights of the Obama Administration. The number of people found to be inadmissible to the US was down from the high in FY2016, too. Thus, so far it does not appear that the Trump administration has greatly restricted immigration.
Conclusion
Throughout much of the 19th century, anyone who had the means to do so could immigrate to the US. In practice, this meant that mostly Europeans, who could afford to do so, and some Asians, especially Chinese and later Japanese, came to the US in droves. Yet, this changed at the end of the 19th and early 20th centuries, first with the exclusion of Asian immigrants and later with the exclusion of most Southern and Eastern Europeans. Since then, the US has reopened immigration some—and immigrants from new parts of the world have come to the US—but not to the degree of the 19th century.
What accounts for the changes in immigration policy? I argue that business support for immigration has declined, first with the Americanization of the US economy, and then with globalization. Greater ability to trade goods across the world has led to the closure of many firms that once employed immigrant workers. The ability to move production overseas has allowed firms to take their capital to labor rather than bring labor to capital. Finally, increased automation and productivity has allowed the remaining firms to do more with fewer workers. As business support has declined, it has opened up space for anti-immigrant groups and led to increased restrictions.
In the last few years we have seen push back against the free trade and firm mobility that has led to reduced support for immigration, both in the US and many other developed countries. Is this declining support for trade having an effect on immigration policy? For the moment, it seems that it is not for several reasons. First, at least in the US tariffs and other trade measures seem to be relatively ineffective at increasing exports or decreasing imports. It is perhaps the case that global supply chains are very difficult to unravel. Second, we have seen very little wage growth associated with these trade protections in the last few years. This could be because there is still potentially slack in the labor market or because of automation, which means that firms can produce as much as they did before with fewer workers. Finally, for all of Trump’s talk on immigration, his administration has not, given the data available, dramatically decreased immigration.
While we have not seen the breakdown in the relationship between immigration and trade as yet, there are at least three ways that it may breakdown in the future. The first is automation. If automation continues to replace human workers, it is possible that trade protections in the future will lead to much more production but few jobs. Second, with the greying of the populations of the developed world, there is increased need for care workers. Most care work cannot be done overseas or automated at this point, nor is it likely to be in the future. Thus, it is possible that even with increased trade openness, countries may demand more low-skill immigrant workers to take care of the elderly. Thus far, however, we have seen only limited support for immigration for care work (Peters et al. 2019). Third, young people are increasingly supportive of diversity and immigration (National Immigration Forum 2019). The Trump administration’s most discriminatory and inhumane attacks on migration have sparked a backlash in public opinion as well (Gallup 2019), which may lead to more support for immigration in the future. It is possible that concern over vulnerable populations, especially with climate change, may lead to more open immigration policies, not because of business support, but because of the support of the mass public.