Good Economics for Hard Times

by Abhijit V. Banerjee

  A third, closely related point is that employers may want to reorganize production to make effective use of the new workers, which can create new roles for the native low-skilled population. In the Danish case we discussed above, Danish low-skilled workers eventually benefited from the influx of migrants, in part because it enabled them to change their occupations.31 Where there were more migrants around, more native low-skilled workers upgraded from manual to nonmanual jobs and changed employers. While doing so, they also shifted to jobs with more complex tasks and that required more communication and technical content; this is consistent with the fact that the immigrants hardly spoke Danish when they first arrived and could not be rivals for these jobs. The same kind of occupational upgrading also happened during the great European migration to the United States in the late nineteenth and early twentieth centuries.

  More generally, what this suggests is that low-skilled natives and immigrants do not have to compete directly. They may perform different tasks, with immigrants specializing in tasks requiring less communication and natives in tasks that do. The availability of immigrants may actually encourage firms to hire more workers; the immigrants perform the simpler tasks, and the natives switch to complementary, more rewarding tasks.

  Fourth, another way in which migrants complement rather than compete with native labor is they are willing to perform tasks natives are reluctant to carry out; they mow lawns, flip burgers, attend to the needs of babies or sick people. So when there are more migrants, the price of those services tends to go down, which helps the native workers and frees them to take on other jobs.32 Highly skilled women, in particular, are more likely to be able to go out to work when there are many migrants around.33 The entry of highly skilled women to the labor market in turn boosts demand for low-skilled labor (childcare, catering, cleaning) at home or in the firms they manage or run.

  The effects of migrants will also crucially depend on who the migrants are. If the most enterprising move, they may start businesses that create jobs for the natives. If they are the least qualified, they might have to join the undifferentiated mass that native low-skilled workers will have to compete against.

  Who migrates typically depends on the barriers migrants have to overcome. When President Trump compared the migrants from “shithole countries” to the good ones coming from Norway, he most probably did not know that a long time ago Norwegian immigrants were part of the “huddled masses” Emma Lazarus talked about.34 There is actually a case study of Norwegian migrants to the United States during the age of mass migration, in the late nineteenth and early twentieth centuries.35 At the time, there was nothing to stop migration, other than the price of passage. The study compared the families of migrants to the families where nobody migrated. It found migrants tended to come from among the poorest families; their fathers were substantially poorer than average. So, by one of the cute ironies historians (and economists) delight in, Norwegian migrants were exactly the kinds of people Trump would instinctively prefer to keep away. In his eyes, they would have been the “shithole people” of their day.

  In contrast, those who migrate out of poor countries today need to have the money to afford the cost of travel and have the grit (or the advanced degrees) required to overcome a system of immigration control typically loaded against them. For this reason, a lot of them bring exceptional talents—skills, ambition, patience, and stamina—that help them become job creators, or raise children who will be job creators. A report by the Center for American Entrepreneurship found that, in 2017, out of the largest five hundred US companies by revenue (the Fortune 500 list), 43 percent were founded or co-founded by immigrants or the children of immigrants. Moreover, immigrant-founded firms account for 52 percent of the top twenty-five firms, 57 percent of the top thirty-five firms, and nine of the top thirteen most valuable brands.36 Henry Ford was the son of an Irish immigrant. Steve Jobs’s biological father was from Syria, Sergey Brin was born in Russia. Jeff Bezos takes his name from his stepfather, the Cuban immigrant Mike Bezos.

  And even among those not so special to start with, the fact of being an immigrant, in a foreign location, without the social ties that make life richer but also impose limits on the single-minded pursuit of one’s career, can liberate one to try something new and different. Abhijit knows of many middle-class Bengali men who, like him, had never washed their own dishes before leaving home. But, finding themselves short of money and long on time in some British or American town, they ended up bussing tables in a local restaurant and discovered they quite liked doing something more hands-on than the white-collar job they had imagined for themselves. Perhaps the reverse happened to the Icelandic would-be fishermen who, thrown into an unfamiliar place where many more people were going to college, decided it might not be such a bad idea after all.37

  So one very big problem with the supply-demand analysis applied to immigration is that an influx of migrants increases the demand for labor at the same time it increases the supply of laborers. This is one reason why wages do not go down when there are more migrants. A deeper problem lies in the very nature of labor markets: supply-demand is just not a very good description of how they really work.

  WORKERS AND WATERMELONS

  Traveling around Dhaka, Delhi, or Dakar in the early morning, you will sometimes notice groups of people, mostly men, crouching on the sidewalks near important crossings. They are job seekers, waiting to be picked up by someone who needs them for work, often in construction.

  For a social scientist, what is striking, however, is how rare these physical labor markets are. Given there are nearly twenty million people in the greater Delhi area, one might imagine every street corner would have such an assemblage. In fact, one has to look around to find them.

  Signs advertising jobs are also relatively rare in Delhi or Dakar. There are lots of ads on websites and employment portals, but most of those jobs are well out of the reach of the average rural goatherd. By contrast, in Boston the subway is full of announcements for job opportunities, but the ads challenge prospective employees to solve some seemingly impossible riddle to prove their intelligence. They want workers but they don’t want to make it too easy for them. This reflects something very fundamental about labor markets.

  Hiring is different from buying, say, watermelons in a wholesale market, for at least two reasons. One is that the relationship with a worker lasts a lot longer than the purchase of a bag of watermelons; you can switch suppliers next week if you don’t like the melon you got. But even where the laws don’t make it difficult to fire a worker, firing is unpleasant at best, and potentially dangerous if the disgruntled employee becomes enraged. Therefore, most firms will not hire just anyone willing to work for them. They worry whether the worker will show up for work on time, whether the work will be up to snuff, whether they will fight with their colleagues, insult an important client, or break an expensive machine. Second, the quality of a worker is harder to judge than that of watermelons (which professional watermelon sellers are apparently very good at assessing38). Despite what Karl Marx had to say, labor is no ordinary commodity.39

  Firms therefore need to put in some effort to know whom they are hiring. In the case of more highly paid workers, this means they spend time and money on interviews, tests, references, and so forth. This is costly both for the firms and the workers, and seems to be universal. In Ethiopia, a study found that just applying for a midlevel clerical job took several days and repeated journeys. Each application cost the would-be applicant a tenth of the monthly wage he would earn and had a very low probability of leading to a hire, one reason why few people applied.40 For this reason, in the case of lower-paid workers, firms often skip the interview and rely on the recommendation of someone they trust. Relatively few firms hire those who just walk in and ask for a job, even if they say they would accept a lower wage. This of course flies in the face of the standard supply-demand framework. But it is too costly to be put in a position where the employer might want to get rid of a wo
rker. In a striking example, researchers trying to find firms in Ethiopia willing to randomize whom they hired, approached over three hundred firms before they found five that were willing to join the experiment.41 These were jobs where no specific skills were needed, but the firms still wanted to retain some control over whom they hired. Evidence from other studies in Ethiopia suggests 56 percent of firms insist on work experience even for blue-collar jobs,42 and it is also common to ask for a referral from an employer.43

  This has several important implications. First, established workers are much more secure from competition from newcomers than a pure supply-demand model would have us believe. Their current employer knows them and trusts them; incumbency is a huge advantage.

  From the point of view of a migrant this is bad news. To make matters worse, there is a second implication. Think of what an employer can do to punish a worker who is not performing; at worst he can fire the employee. But firing will only be adequate punishment if the job pays enough for the worker to really want to keep it. As the future Nobel Prize–winner Joe Stiglitz pointed out many years ago, firms would not want to pay their workers the minimum the workers would accept, precisely to avoid being in the position captured by that old Soviet joke: “They pretend to pay us, we pretend to work.”

  This logic says that the wage the firm must pay to get workers to work typically has to be high enough that being fired actually hurts. This is what economists call the efficiency wage. As a result, the wage difference between what firms pay their established workers and what they would need to pay a newcomer may not be very large, because they cannot risk the consequences of paying a newcomer too little.44

  This makes the incentive to employ a prospective migrant even weaker. Moreover, employers are also reluctant to have large differences in wages within their establishments, for fear of lowering morale. Evidence suggests that workers hate inequality within firms, even if the inequality is related to productivity, at least when the link between pay and productivity is not immediately obvious and transparent.45 And unhappy workers do not make for a productive workplace. This contributes to why native workers are not quickly replaced by cheaper immigrants.

  This discussion fits nicely with another finding from the Czech migration study mentioned before: job losses for natives were not actually losses; they were, rather, lower gains (compared to regions of Germany where Czechs did not go).46 German firms did not replace their existing staff with Czech migrants. Those already employed in Germany still had the benefit of familiarity. What happened was that instead of hiring new native workers whom they did not know, German firms sometimes hired Czechs whom they also did not know.

  The view that there is not much scope for migrants to get the jobs natives already hold, even by offering to do them for lower wages, also helps us understand why immigrants often end up in jobs natives do not want, or in cities no one wants to go to. There, they are not taking jobs from anyone; those jobs would remain unfilled if there were no migrants willing to take them.

  THE SKILLED SET

  So far we have been talking of the impact of unskilled migrants on natives. But even those who oppose unskilled migration are usually in favor of skilled migrants. Many of the arguments we made to explain why low-skilled migrants do not compete with low-skilled natives do not apply to skilled ones. For one, they are typically paid much more than the minimum wage. There may not be a need to pay them an efficiency wage because their jobs are exciting, and getting a chance to do them and doing them well would be its own reward. Therefore, there is paradoxically more scope for a skilled migrant to undercut the wages of the natives. Second, for skilled workers, the employer cares relatively more about the exact skill set of the person being hired than about the applicant’s personality or reliability. Most hospitals hiring a nurse, for example, will focus primarily on whether the applicant meets the legal requirements for the job (in particular, whether they have taken and passed the nursing board exam). If a foreign-born nurse with the right certification is available for less, the hospital has little reason not to go for that nurse. Moreover, no one hires such workers without a series of interviews and tests, putting unknown workers on the same footing as familiar or connected ones.

  Therefore, it is no surprise that in the United States one study finds that, for every skilled, qualified foreign nurse employed in a city, there are between one and two fewer native-born nurses.47 In part this is because native-born students facing competition from nurses born and educated abroad are unwilling to sit for the nursing board exam in their states.

  Therefore, despite widespread support for it, including from people like President Trump, the immigration of skilled workers is more of a mixed bag from the point of view of its impact on the domestic population. It helps low-skilled natives, who benefit from cheaper services (most doctors who serve the poorest corners of the United States are migrants from the developing world) at the cost of worsening the labor market prospects of the domestic population with similar skills (nurses, doctors, engineers, and college teachers).

  WHAT CARAVAN?

  The myths about immigration are crumbling. There is no evidence low-skilled migration to rich countries drives wage and employment down for the natives; nor are labor markets like fruit markets, and the laws of supply and demand do not apply. But the other reason immigration is so politically explosive is the idea that the numbers of would-be immigrants are overwhelming, that there is a flood of strangers, a horde of foreigners, a cacophony of alien languages and customs waiting to pour over our pristine monocultural borders.

  Yet, as we saw, there is simply no evidence the hordes are waiting for a chance to descend on the shores of the United States (or the United Kingdom or France) and need to be kept out by force (or a wall). The fact is that unless there is a disaster pushing them out, most poor people prefer to stay home. They simply aren’t knocking on our door; they prefer their own countries. They don’t even necessarily want to move as far as their local capital city. People in rich countries find this so counterintuitive that they refuse to believe it, even when faced with the facts. What explains it?

  WITHOUT CONNECTIONS

  There are many reasons why people don’t move. All the things that make it hard for new immigrants to compete with long-term residents for jobs also discourage them from moving. For one, as we saw, it is not easy for an immigrant to find a decent job. The one exception is where the employer is a relative or a friend, or a friend of a friend, or at least a co-ethnic: someone who either knows or at least understands the migrant. For that reason, migrants tend to head to places where they have connections; finding a job is easier and they have help to land on their feet in the city. Of course, there are all kinds of reasons why the employment prospects of migrants from the same location will be correlated over time; for example, if a village produces great plumbers, both recent and previous generations of migrants will be employed, and employed in plumbing. But the pull of kinship is stronger. Kaivan Munshi, a professor at the University of Cambridge, and perhaps not coincidentally a member of the small and very tightly connected community of Zoroastrian Indians otherwise known as Parsis, demonstrated that Mexican migrants explicitly seek out people they might know.48

  He observed that, regardless of opportunity in the United States, bad rains (disasters) have pushed people out of Mexico. When the rains failed in a particular village, a group of people left to seek other opportunities. Many of them ended up in the United States, with the result that a subsequent migrant from the same village would have connections in the US who were securely employed and able to help him or her find a job. Kaivan predicted that if one compares two villages in Mexico that have the same weather this year, but one of them had a drought several years ago (causing some villagers to emigrate) while the other did not, it will be easier for a resident of the village with the past drought to find a job (and also to find a better job) than for the resident of the village without the past drought. He expected to see more migrants, more em
ployed migrants, and better-paid migrants. This is exactly what the data showed. Network connections matter.

  The same applies to the resettlement of refugees; the ones most likely to find employment are those sent to a place with many older refugees from the same country.49 Those older refugees usually do not know their new countrymen, but they still feel compelled to help.

  Connections are obviously useful for those who have them, but what happens to those who don’t? They will clearly be at a disadvantage. In fact, the presence of some people who come with recommendations can ruin the chances for everyone else. An employer used to workers coming with recommendations is likely to be suspicious of anyone without one. Knowing that, anyone who can get a recommendation would rather wait to get it (maybe some connection to a prospective employer will emerge; maybe a friend will start a business), and only those who know no one will ever recommend them (perhaps because they are actually not good workers) will go around knocking at doors to find a job. But then the employer would be right in refusing to talk to them.

  The market in this situation is unraveling. In 1970, George Akerlof, another future Nobel laureate, but then just a fresh PhD, wrote a paper, “The Market for ‘Lemons,’” in which he argued that the market for used cars might just shut down because people have an incentive to sell off their worst cars. That sets off the kind of self-confirming reasoning we saw in the case of newcomers to the labor market; the more suspicious buyers become of the old cars being sold, the less they will want to pay for them.50 The problem is the less they want to pay, the more the owners of good used cars will want to hold on to them (or sell their cars to friends who know and trust them). Only those who know their car is about to collapse will want to sell on the open market. This process by which only the worst cars or the worst employees end up on the market is called adverse selection.51