In response to record numbers of illegal border crossings and the security fears triggered by the 9/11 attacks, over the past two decades the United States has steadily increased its efforts to secure its borders against illegal immigration. The number of U.S. Border Patrol agents has risen from fewer than 3,000 to more than 20,700; nearly 700 miles of fencing have been built along the southern border with Mexico; and surveillance systems, including pilotless drones, now monitor much of the rest of the border. In a speech in El Paso, Texas, in May, U.S. President Barack Obama claimed that the United States had "strengthened border security beyond what many believed was possible." Yet according to spring 2011 Rasmussen poll, nearly two-thirds of Americans think the border is no more, or even less, secure than it was five years ago. Some administration critics claim that the United States' frontiers have never been more porous.

This contradiction stems in part from the fact that the Department of Homeland Security (DHS) has never clearly defined what border control means in practice. A secure border cannot mean one with no illegal crossings -- that would be unrealistic for almost any country, especially one as big and as open as the United States. On the other hand, the borders cannot be considered secure if many of those attempting to enter illegally succeed. Defining a sensible middle ground, where border enforcement and other programs discourage many illegal crossings and most of those who try to cross illegally are apprehended, is the challenge.

Unfortunately, the U.S. government has failed to develop good measures for fixing goals and determining progress toward them. Since 2005, the DHS has reported how many miles of the country's land borders are under its "operational control," but it has done so without having clearly defined what that standard means and without providing hard data to back it up. The lack of sound measurement has left the administration touting its efforts rather than their results: during a press conference in 2010, Obama noted, "We have more of everything: ICE [Immigration and Customs Enforcement], Border Patrol, surveillance, you name it. So we take border security seriously."

It is no wonder, then, that many critics dispute the president's claims. In Congress, the result has been a stalemate over border policy that has stymied progress on every other aspect of immigration reform. Obama's opponents, led by longtime supporters of immigration reform, including Republican Senators John McCain, Lindsey Graham, and Jon Kyl, have demanded that the borders be secured before any other aspect of immigration reform is addressed. Last year, for example, Congress approved almost unanimously an additional $600 million for border enforcement, on top of the $17 billion annual budget. But every other immigration proposal -- including the DREAM Act, a bill that would grant residency to the children of illegal immigrants who attend a U.S. college for two years or join the military; new programs for temporary agricultural workers; and increased quotas for high-skilled immigrants -- has been held up because of congressional demands that the border be secured first.

To move the debate on border security beyond politically driven speculation to a more serious consideration of how much enforcement is needed, and at what cost, the Obama administration must develop effective ways to measure progress on border security and then inform Congress and the public regularly about it.


For most of its history, the United States has known little about the people who attempt to cross its borders illegally. Beyond its yearly count of apprehensions, most of which occur along the U.S.-Mexican border, the Border Patrol reports very little data to the public. In 1986, the agency recorded the largest annual number of arrests -- 1,693,000 -- as immigrants flooded to the border hoping to take advantage of a new immigration bill that offered legal residency to many who were already illegally in the United States. In 2000, the number of arrests was almost as high, driven by the strong U.S. economy and rising numbers of young workers looking for higher wages in the United States. Arrests have dropped sharply since then. Only 463,000 were made in 2010, the lowest number since 1972.

On its own, the number of apprehensions gives only a hazy picture of the number of immigrants actually trying to sneak across the border, since a given individual may be caught multiple times or may enter without ever being caught. Moreover, arrest figures can highlight patterns of illegal immigration but cannot explain why they change. In the early 1990s, for example, Border Patrol arrests near the Californian-Mexican border averaged nearly 600,000 annually. After the government strengthened enforcement there in the mid-1990s, apprehensions fell by half in California but soared in Arizona, indicating that the illegal traffic had merely shifted east. (Today, Arizona remains the largest illegal-entry corridor into the United States. This helps explain the political backlash against immigration that led to the state's 2010 adoption of the controversial law known as SB 1070, which increases the state police's powers to arrest suspected illegal immigrants.)

Without data to supplement annual arrest figures, it is hard to say whether the past decade's falling apprehension numbers are a victory for enforcement or simply the result of a weaker U.S. economy that is no longer a magnet for immigrant workers. For its part, the DHS claims that effective enforcement explains the recent drop in the number of apprehensions. Yet a decade ago, the DHS' predecessor agency, the Immigration and Naturalization Service, claimed that better enforcement accounted for the peak in the arrest numbers. A measure that indicates success whether it rises or falls is not very useful without further information.


On its Web site, the Border Patrol states that "although the Border Patrol has changed dramatically since its inception over 80 years ago, its overall mission remains unchanged: to detect and prevent illegal entry of aliens into the United States." Currently, the DHS claims that nearly 900 out of the 1,969 miles of the U.S.-Mexican border are under the Border Patrol's control, but it has not presented hard data to support that claim.

A good place to start is estimating apprehension rates, that is, the number of arrests compared to total illegal crossings. With all the manpower, fencing, and surveillance that have been deployed along the U.S.-Mexican border, one would expect the U.S. government to be catching a higher percentage of illegal immigrants today than ever before. But the government currently does not report that rate.

Reasonable estimates can certainly be compiled. Since 1997, the Border Patrol has made internal calculations of "known illegal entries." The numbers come from close observation conducted by individual agents, who collect information about footprints and other physical evidence of human traffic across the U.S.-Mexican border, and from images from surveillance cameras. However, these Border Patrol estimates are rarely cited publicly, and when they are, they are used in a highly selective fashion. For example, a recent Government Accountability Office report stated that 90 percent of known illegal entrants in one sector, in Yuma, Arizona, were either apprehended or deterred from entering in 2009. But Yuma has seen a huge enforcement buildup in recent years, so very few immigrants attempt to cross there.

To be sure, this measure is almost certainly higher than the true apprehension rate, since known illegal entrants do not include those who successfully cross unnoticed. In contrast, another component of the DHS, the U.S. Coast Guard, has been calculating and reporting an apprehension rate based on known illegal entries at sea since the mid-1990s. It estimates that the apprehension rate at sea is about 60 percent.

Academic studies have also offered estimates of the apprehension rate. In 1994, Thomas Espenshade of Princeton's Office of Population Research developed a model that uses fingerprint data to identify individuals that the Border Patrol has caught trying to sneak across the border more than once. According to his research, evidence from those caught multiple times can be used to estimate with some accuracy the likelihood that an immigrant will be arrested on any given trip to the United States. He concluded that between 1977 and 1988, when there were only one-sixth as many Border Patrol agents as there are today, the apprehension rate was about 30 percent. This finding echoes studies by Princeton University's Mexican Migration Project, which since the 1980s has surveyed households in Mexican towns that have high emigration rates. They, too, suggest that the apprehension rate of those trying to cross illegally into the United States was around 30 percent between 1965 and 2005. Recent data from surveys of immigrants suggest that the rate may have risen in recent years.

If the estimates based on known illegal entries are too high, these are almost certainly too low, because they do not take into account immigrants who give up trying to cross after being caught. Taking deterrence into account, the true apprehension rate today for the entire southwestern border is likely between 40 and 50 percent, and it may have risen in recent years due to increased enforcement resources. The DHS is reluctant to publish a figure, however, probably because it varies greatly among different border sectors and because the data are better for some sectors than others. Although a high apprehension rate is a positive indicator, it is still insufficient as a measure of success. As with all law enforcement, the real aim of border enforcement is to deter illegal acts. Effective enforcement is impossible if lawbreaking is the norm. Even a very high apprehension rate would be less than desirable if many people kept trying.


There are two types of deterrence relevant to border security: behind-the-border deterrence, in which enforcement discourages would-be immigrants from ever trying to cross illegally, and at-the-border deterrence, in which those who have been caught crossing the border at least once are deterred from trying to do so again. In an ideal system, the first kind of deterrence would work well enough to obviate the need for the second kind. Assessing behind-the-border deterrence is difficult, however, since it requires information about both people who decided to immigrate illegally and those who decided not to for fear of being caught.

The long-standing consensus of most academic research on illegal immigration, from such respected scholars as Douglas Massey of Princeton University and Wayne Cornelius of the University of California, San Diego, is that enforcement has little behind-the-border deterrent effect. Instead, this research argues that attempted entries are largely driven by economics; when the U.S. economy is strong, especially in sectors that employ many illegal immigrants, such as construction, the numbers rise. This has certainly been the historical pattern: total apprehension figures dropped during the recessions of 1981-82, 1990-91, and 2001-2.

Yet there is some evidence that deterrence is growing. The latest decline in apprehensions began well before the 2008 recession and has yet to be reversed, even as the U.S. economy is recovering. But without clearer models of illegal immigration, the administration cannot isolate the role of enforcement. The number of illegal crossings will likely rise again as the U.S. economy continues to recover, but since the current metrics are so poor, it will be impossible to tell whether that increase should be attributed to better economic conditions in the United States or to deficiencies in border security.

At the moment, the only available measure of at-the-border deterrence -- from surveys of Mexican immigrants -- is discouraging. It suggests that in the last decade, 90-98 percent of the Mexican nationals who tried to enter the United States illegally were ultimately successful. The odds of apprehension on any given journey have likely increased, but persistence continues to pay off. This appears to be due to the involvement of professional smugglers. Evidence suggests that half of all illegal crossers hired smugglers in the 1970s and that this proportion has risen significantly since. Data on the average fee paid to smugglers, which are collected by the immigrant surveys, show that, adjusted for inflation, the fee doubled between the early 1990s and 2005. In part, these increasing fees may explain both the apparent increase in behind-the-border deterrence and the minimal effectiveness of at-the-border deterrence: some immigrants cannot afford smugglers, so they stay at home, and those who decide to pay a smuggler are determined to try until they succeed. (Immigrants typically pay smugglers half the fee in advance and half on their successful entry, so smugglers will guide the immigrants on multiple attempts.) Smuggling plays a critical role in facilitating illegal entry into the United States, and the DHS needs to collect, analyze, and report data on smuggling operations more systematically.

The Border Patrol has long recognized the importance of deterrence at the border. But for decades, it simply returned most of the Mexican immigrants it apprehended to just over the border, only to see them try to cross again and again. It has recently begun imposing penalties, such as jail time or a flight to the interior of Mexico. There is not yet good data, however, on whether these new methods are increasing deterrence.

If the goal of border enforcement is to prevent successful illegal entry into the United States, the government will need to devise better ways to track those who overstay their visas or are smuggled in through legal ports of entry. Ultimately, the U.S. government needs to generate reasonably accurate measures of the total inflow of illegal immigrants in order to develop the most effective policy responses. Measuring illegal immigration may seem inherently difficult, but in 2000 and 2001, the government explicitly set the twin goals of reducing both illegal immigrant inflows and the total number of unauthorized immigrants in the United States. And those numbers are routinely estimated by experts both inside and outside the government. After the DHS was formed in 2002, the regular measuring of progress toward these goals stopped, but it is clear that taking such measurements is possible if the political will exists.


Systematically reported apprehension rates and deterrence data would give Congress better information to decide whether the money spent on border security is wisely allocated. Border security budgets have ballooned over the past decade, but the results are unclear. Right now, questions about whether current spending is adequate to control the border and, if not, how much would be are largely unanswerable.

With better measures, the government could experiment with new programs. For example, increased numbers of legal work opportunities would likely reduce incentives to cross illegally. But unless Congress had some confidence that this could be measured and verified, it would be reluctant to authorize the program. If the government could better track illegal crossings, it could test pilot programs of new initiatives as an intermediate step.

Similarly, improved measurement could help determine the relative effectiveness of various enforcement tools. The DHS has no real understanding of how much border control improves when the number of Border Patrol agents, sensors, or miles of fencing increases. It also has no understanding of the relative effectiveness of border enforcement as compared to efforts to prevent employers from hiring illegal workers.

Signs that the DHS and Congress are finally starting to seek more reliable border security metrics are mixed. At a February 2011 hearing before the House Homeland Security Subcommittee on Border and Maritime Security, Michael Fisher, the Border Patrol chief, noted that "operational control is not in and of itself an assessment of border security." The Border Patrol has said that the agency is working on a new set of performance metrics that will improve the assessment of border security. However, at a hearing in May, DHS Secretary Janet Napolitano instead promised new measures of success focused on "overall security and quality of life along the entire border region." Such assessments would miss the point: it would be as if local police forces measured property values rather than crime rates. Better border enforcement will certainly improve the quality of life along the border, but that is a consequence of a job well done, not the job itself. The DHS needs to develop and report measures that directly address its mission, and the government will need to establish clear targets and genuine metrics for progress. Without them, the policy debate will remain mired in unfounded claims and immeasurable goals.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • EDWARD ALDEN is Bernard L. Schwartz Senior Fellow at the Council on Foreign Relations. BRYAN ROBERTS is Senior Economist at Nathan Associates and was Assistant Director of Borders and Immigration in the Office of Program Analysis and Evaluation at the U.S. Department of Homeland Security in 2010.
  • More By Edward Alden
  • More By Bryan Roberts