The biggest immigration debate of this year in the US has been what to do about the rise in migration pressure at the Southwest border. That pressure comes mostly from the “Northern Triangle” of Central America: Guatemala, Honduras, and El Salvador.
Effective policy for managing this pressure needs to rest on three big facts that have received much too little attention in the debate so far:
An imminent demographic shock in the region will ease migration pressure
Visa policy is critical to strengthen law enforcement
The effect of aid on migration is more complex than many believe
1. Central America is falling off a demographic cliff—so migration will slow
The Northern Triangle has recently begun to fall off a demographic cliff. There will be fewer youths entering the labor market in the region in years to come than since the 1950s. In roughly a decade, migration pressure is likely to fall sharply as a result. Much of today’s pressure will naturally ease.
The best evidence comes from the historical experience of migration from Mexico. From about 1970 to about 2010, there was a very large wave of unauthorized immigration to the US from Mexico. That wave can be almost entirely explained by the surge of youths into the Mexican labor market that happened about a decade earlier, caused by long-term changes in Mexico’s child mortality and fertility as the country developed.
Border apprehensions of Mexican migrants rose and fell in tandem—with a delay. Below is the change in the population of young working-age people in Mexico 1955–2005 (on the left axis) compared to US border apprehensions of Mexican migrants 1970–2018 (on the right axis).
Source: Demographic data from UN Economic Commission for Latin America, US border apprehensions of Mexican citizens from INS Statistical Yearbook and DHS Yearbook of Immigration Statistics
This striking pattern would not make sense unless one of the key reasons that so many Mexicans came to the United States—often seeking work—was job competition at home. When there were fewer youths entering the labor market, fewer sought work abroad.
Statistically, this relationship is extremely tight. Knowing only the percentage change in the number of young workers in Mexico in each year 1955-2005, you can predict 83 percent of all changes in US border apprehensions of Mexicans in each year, 13 years later, across the whole period of rise and fall, 1970–2018.
Think about that. Knowing just one big change in Mexico, you can statistically account for almost the entire wave. Of course, many other forces were exerting a “push” and “pull” on unauthorized migration from Mexico during that period. These include the 1982 economic crisis in Mexico, the 2001 and 2008 economic crises in the US, the NAFTA accord, and big increases in US border enforcement. All of those mattered, as research has shown persuasively. You can see some of them clearly as important blips in the above figure, like the spike in apprehensions after Mexico’s crisis of 1982.
Think about that. Knowing just one big change in Mexico, you can statistically account for almost the entire wave.
But in the big picture, none of those causes can compete with demographic change. Consider ten young workers, thinking about attempting to migrate. Stepped-up border enforcement could exert a big effect if it deterred, say, three of those ten. But that effect would be swamped by a demographic shift implying that nine of those ten were not there in the first place. This was a major reason for the collapse of unauthorized immigration by Mexicans. That is why economists Gordon Hanson and Craig McIntosh, using only data available 19 years ago, were able to accurately predict today’s low levels of Mexican migration.
And just the same thing is about to happen in the Northern Triangle. The graph below shows that the very same youth surge has been happening in the Northern Triangle, about 30 years after Mexico. In all three countries, the annual growth in the number of young workers has been peaking, while we’ve seen high unauthorized migration from the Northern Triangle. And that youth surge has just started to collapse:
Source: Demographic data from UN Economic Commission for Latin America
The numbers to the right of 2019 are forecasts, by the United Nations. (These are much more reliable than other kinds of forecasts. Out to the year 2034, the people represented above are already born.)
In El Salvador, the number of young workers is actually shrinking; growth is already negative. The same thing will happen in Honduras in a few years. That has never happened in Mexico. And the sharp slowdown in Guatemala will happen twice as fast as it happened in Mexico.
In other words, this demographic cliff suggests that high emigration from the Northern Triangle is a temporary phase. The Great Mexican Emigration, as Hanson and McIntosh call it, had a beginning and an end. We have been living through the Great Northern Triangle Emigration, and it is about to end—for the same reasons. Policies to manage high migration pressure from the Northern Triangle should be designed not as a panicked response to unlimited migration pressure, but as temporary tools to manage a transitory phenomenon.
2. Lawful migration channels are a necessary tool—but far from enough
The second missing piece in the policy debate on the Northern Triangle is lawful migration channels, especially work visas. Work visas are often seen as the antithesis of border enforcement. In fact, work visas are a powerful tool for border enforcement.
Here again, the best historical lessons come from real experience with Mexico. Unauthorized immigration by Mexicans has simply collapsed, falling 91 percent since the year 2000. As we’ve seen, demographic change was a critical reason for that. Another was work visas.
That collapse of Mexican unauthorized migration happened exactly when work visas to Mexicans exploded by a factor of ten. Since 2000, the number of US admissions of Mexicans on seasonal work visas (H-2A and H-2B visas) has shot up from about 55,000 to over 550,000 each year. Here is a picture of Mexico’s massive flip from unauthorized to authorized migration:
Source: US border apprehensions of Mexican citizens and US admissions of Mexican citizens on H-2A, H-2B (including variant H-2R) visas estimated based on INS Statistical Yearbook and DHS Yearbook of Immigration Statistics.
Here again, this is an extremely tight statistical relationship. Knowing only the percentage change in Mexican entries on lawful work visas each year, you can still predict 88 percent of the declines in apprehensions of Mexicans since 2000. Every one percent rise in authorized admissions was associated with a one percent fall in border apprehensions. (How can both demographic change and visas explain so much of the fall in apprehensions? Because they are not separate causes, they are intimately related. For example, reduced availability of unauthorized workers due to demographic change has driven many reluctant employers to bear the large costs of sponsoring work visas.)
This certainly does not mean that work visas by themselves caused the fall in apprehensions. Economist Pia Orrenius has shown that stepped-up border enforcement is strongly associated with lower unauthorized immigration too, as has a stack of other research, even as deep long-term research by sociologist Doug Massey and colleagues has demonstrated the limits of an enforcement-only approach. The apprehension rate (per crossing attempt) at the border did not change during the whole first decade after 2000, during which apprehensions fell about 80 percent. Greater enforcement simply cannot be the primary explanation for the collapse, by itself, but neither can greater work visas.
Instead, this evidence suggests that the effectiveness of increased enforcement depended on the existence of lawful channels toward which enforcement could divert migration pressure. That is: the history of the US-Mexico border suggests that enforcement and lawful channels together were effective in reducing unauthorized migration—but only in combination.
This matters for policy in the Northern Triangle today, because, in that debate, lawful channels have been nearly off the radar. Northern Triangle access to US seasonal work visas has been very low and stagnant in recent years, even as Mexico’s access has doubled relative to its population.
Source: H-2A, H-2B, and H-2R visas granted each year by citizenship from US State Department, Report of the Visa Office. Population from the United Nations World Population Prospects 2019 release.
Last year, the US granted Mexico 1.9 seasonal work visas per capita. It granted Guatemala only 0.4 seasonal work visas per capita. For Honduras and El Salvador, it was just 0.1. One of the fastest and most effective tools against unauthorized migration from the Northern Triangle would be to raise the region’s access to US work visas. It would take an order-of-magnitude increase in visas per capita to make a substantial difference. That would not require a vast change in the allocation of US work visas, because the Northern Triangle is relatively small: only 23 percent of Mexico’s population.
Canada has done this in partnership with Guatemala for the last 16 years. The US has not even tried.
This is not something the US government could simply do by fiat, because US seasonal work visas are employer-led. It would require active facilitation of relationships between US employers and workers from the Northern Triangle. That is clearly feasible. Canada has done this in partnership with Guatemala for the last 16 years. The US has not even tried.
Another effective tool barely on the radar in Washington is work visas further south, within the region. It’s barely known in the United States, but, in 2012, Mexico created a temporary work visa for individual breadwinners from Guatemala (and Belize) to work lawfully in its own southern-border states. The program is in its infancy but already large enough to affect flows further northward; in fiscal year 2018, more than 7,500 Guatemalans worked in Mexico on that visa, while the US border patrol apprehended about 43,000 Guatemalan single adults. If the program continues to expand, it could come to offer a meaningfully large alternative for labor migrants to work lawfully within the region. The US could support both Mexico and Guatemala in promoting use of the visa and creating employment in the Mexican border states it covers.
These useful, pragmatic tools are mostly ignored. US policy today is fighting to reduce unauthorized migration from the Northern Triangle with one hand tied behind its back. A few observers have been highlighting the critical importance of US work visas, including David Bier and Matthew La Corte. Tools to shape mobility within the region have been highlighted by Andrew Selee and Ariel Ruiz, along with leading Mexican researchers Silva Giorguli and Claudia Masferrer. The Bipartisan Policy Center has advanced a range of proposals for effective US cooperation with Mexico and with Guatemala. The Center for Global Development, too, convened a bipartisan, binational commission to build one such proposal.
Such tools are barely mentioned in politicians’ proposals. The US government’s enforcement-first-and-last approach leaves out one of the most effective tools against unauthorized immigration it has used in the past.
3. The effect of aid on migration is complex—because so are migrants’ motives
The US has made its aid to the Northern Triangle conditional on local governments interdicting migrants. But if aid in the region is going to affect migration sustainably, it must be guided by the reasons people move. Northern Triangle migration is often portrayed as caused almost entirely by violence, poverty, or family ties. A big reason is that these three “causes” match up neatly with legal categories of US entry: asylum-seeking, work visas, or family sponsorship visas. The immigration law system is designed to make each individual migrant fit just one of these categories.
The reality is much more complicated. Migration from the Northern Triangle is driven by all of these forces at once. The flow contains lots of individuals in all three categories, and all three forces often motivate the same person.
This was the clear result when one of us (Clemens) studied the statistical drivers of unaccompanied child migration from the Northern Triangle. That research used confidential individual-level data on every unaccompanied child from the region apprehended at the US border between 2011 and 2016, linking them to homicide rates and economic conditions in the local area where each child originated.
The study found that violence was an important driver of child migration. Each ten additional homicides in a child’s municipality of origin caused at least three (and probably six) additional unaccompanied children to be apprehended at the US border. But economic conditions at the origin explained roughly as much of the volume of child migration as violence did. And, critically, violence and economic explanations affected each other. For example, children from the very poorest places—where smugglers’ high fees are out of reach—were less likely to react to violence by migrating.
That is, the effect of violence on migration depends heavily on context, including economic context. Here is a map of the Northern Triangle, showing where additional violence was the most (red) and least (green) likely to spur additional child migration. Violence produced more child migration, on average, in relatively richer parts of the region:
Source: Clemens (2017)
The major implication: better economic opportunity in the region is unlikely to reduce migration—by itself, anytime soon. The places in the Northern Triangle with the least child migration have not been the richest places, but the poorest: departments like Alta Verapaz in Guatemala and Gracias a Dios in Honduras.
You can see this most clearly in the simple correlation between the intensity of child migration from each municipality of the region, and the poverty rate in that municipality:
Source: Clemens (2017). Average across municipalities. Shaded area shows 95% interval of statistical confidence. UAC data pooled across 2011–2016. Poverty fraction estimated in 2007, at national poverty line of each Northern Triangle country.
That is, less poverty is strongly associated with more child migration from this region. Similar forces drive the migration of families with their children. As families in poverty get a little more disposal income, they often invest that income in migration. That, too, happened in Mexico: that wave of Mexican emigration you saw above happened during a period (1970–2005) poverty fell by roughly half, and real income per capita in Mexico more than doubled, from about US$6,000 to US$13,000 in constant 2011 dollars. Far from quickly halting migration, migration from Mexico went hand-in-hand with development. It does in the Northern Triangle too.
None of this means that the drivers of this migration are predominantly related to security or to economics. Those effects cannot be separated from each other.
Again, none of this means that the drivers of this migration are predominantly related to security or to economics. Those effects cannot be separated from each other.
Think about it like this. Suppose that a girl from a slum of San Pedro Sula, Honduras traveled to the US border earlier this year. Her family made the decision to take her there because her prospects for education and a career are bleak at home, where gangs are the local authority. They are also afraid because some of her classmates have been killed. They decided to make the trip because an uncle in Virginia said they could stay with him while they got on their feet, and helped with their expenses. They were able to pay because their own parents had transcended low-skill agricultural work and emerged into the urban lower middle class.
Is this girl’s migration driven by violence, by economic opportunity, or by family reunification? It is all three of them at the same time, for the same person. Does poverty make the family’s decision more likely? National poverty contributes to the lack of local opportunity for the girl, certainly, but her own family’s emergence from poverty is part of what places migration within their reach. This is why, as poor countries get rich, emigration typically rises at first, only falling later.
These core facts must guide any effective policy toward the region, especially aid policy. Aid can certainly shape some drivers of migration, to a limited degree, as researchers including Sarah Bermeo have pointed out. Aid interventions supporting community-driven anti-gang programs have been shown to reduce violence in the region. Food insecurity, another important driver of emigration, can be mitigated with aid programs. There is some evidence that youth employment can modestly deter emigration in the short run, and aid can help generate such employment, though the effect is often very small.
The ultimate success of development aid is prosperity and poverty reduction, and these things may encourage Northern Triangle emigration for a time—as they did in Mexico. This effect is not at all a reason to abandon or reduce aid to the region. It is a reason not to look to development aid for a migration “quick fix.” Effective policy toward the region requires assistance to community-driven security programs, help with food insecurity, and fostering youth employment. It also requires the enforcement of US law in partnership with regional governments.
But these mainstream policies are not enough. Enforcement works best when complemented by substantial lawful migration channels. And aid for Central American development and security, which are clearly in the US long-term interest, is no easy solution in the shorter term.