President Trump’s decision to end the DACA program and send hundreds of thousands of immigrants back into the shadows is cruel. But lost in the moral argument is the fact that it’s also extraordinarily poor policy for fighting crime and keeping America safe.

The President and his Attorney General, Jeff Sessions, frequently bemoan the epidemic of crime in the United States, often implicitly linking crime to immigration. In fact, violent crime is lower now than at any time since the 1950s. Both legal and illegal immigrants are far less likely to commit violent crimes than native-born Americans. And making more immigrants illegal decreases their trust in police – destroying the ability of law enforcement to gather intelligence on real lawbreakers, thus destroying the successful policing policies responsible for the extraordinary fall in crime in America since the 1990s. Let me break these three points down in detail.

Violent Crime in America is at an Extraordinary Low Point.

A graph of homicide data from the Center for Disease Control shows the stunning decline in violence in America:

You can see that the so-called “largest increase” in homicide rates in 45 years that President Trump likes to talk about at the far right of the graph – that’s what an uptick looks like when violence rates are this low. If one person is killed one year, and two are killed the next, it’s a 100% increase in murder – but few would call it a crime wave – there’s no telling, at this point, whether it’s just a statistical blip or a trend. Yet Trump’s own policies are likely to push toward the latter.

Immigrants Are Less Likely to Commit Violent Crimes than Native-Born Americans

There are few points on which the vast majority of scholars agree. One of them is that immigrants, both legal and illegal, are far more law-abiding than native-born Americans.  The findings are so strong that some scholars argue that part of the steep fall in violence in the 1990s was caused by higher than normal immigration during that period.

The fact that U.S. locales with higher rates of immigration have lower homicide rates is echoed by research conducted by Reid et al., Wadsworth, Ousey and Kubrin, and Stowell et al.  The findings held for Los Angeles in the 2000s, and for gang-ridden San Diego from 1980 to 2000 when immigrants were pouring into the city: As immigrants arrived, homicides fell.  In fact, pretty much the only slightly negative correlation between immigration and crime comes from Jörg Spenkuch, who found that a 10% increase in foreign-born immigrants with poor employment outlooks raises a county’s property crime rate by just over 1%, but causes no rise in violence.

This is no surprise when you know that immigrants themselves are far less likely to commit crimes – especially violent crimes – than native-born Americans. In fact, immigrants started out less violent than native-born immigrants, and have become less and less crime-prone with each census since 1980.  By 2000, native-born Americans were five times more likely to be incarcerated than immigrants. That particularly holds true for less-educated Mexican, Salvadoran, and Guatemalan young men, who are overrepresented among illegal immigrants.  By 2010, more than 10 percent of native-born men aged 18-39 without a high school diploma were incarcerated. The percentage for Central American immigrants? Just 1.7 percent.

Immigration opponents often mention that 22% of federal inmates are not citizens. But since the vast majority of prisoners are in state prisons, that rate represents just 27,505 people likely to be undocumented immigrants among all those sentenced to federal prison in 2014 (the last year statistics were available). Of that number, 20,333, or 74%, had been sentenced for immigration violations such as re-entering the country after being deported. Only nine had been convicted of murder, 50 for assault, and 19 for sexual abuse.

The benefits of immigration on crime extend beyond the immigrants themselves  – it turns out immigrants are also good influences on their peers. A study of 3,000 Chicagoans found that an individual’s chances of being violent grew with each generation’s distance from immigration, so that someone in the U.S. for three generations was twice as likely to be violent as a first-generation immigrant. But Desmond and Kubrini found that young people (always the demographic most prone to violence) who had the good luck to live in neighborhoods with a greater number of first-generation immigrants were less likely to be violent. That means that not only are the immigrants themselves less prone to violence but they also likely reduce the level of violence among non-immigrant youths within their network or community.

It could be that illegal immigrants are afraid of deportation and so refrain from committing crimes, or that legal immigrants walk the straight and narrow due to old habits from authoritarian countries of origin. But since immigrants also tend to have lower rates of anti-social behavior across a whole range of indicators, from juvenile delinquency to substance abuse, it’s more likely that the old stereotype of hard-working, decent people who come here to better their lives is closer to the truth for those who have legal paths in, and the same goes for those who have entered illegally or been brought here as children, as with the DACA program.

Deportation and Immigration-Focused Policing Could Increase Crime

If findings this strong applied to any policing policy, police departments around the country would be clamoring to improve their effectiveness by increasing immigration to decrease violent crime. That’s what evidence-based, objectively minded public policy would essentially dictate.

Instead, President Trump and some of his close advisors tout tough deportation of criminal immigrants as a way to increase safety. A study of President Obama’s stepped-up criminal deportations at first seems to support Trump’s approach: crime declined by 4% as Obama’s so-called “Secure Communities” program detained 250,000 immigrants with criminal records (more than the total existing federal prison population) from 2008 to 2012, But once researchers (in peer-edited publications) took into account the general decline in U.S. crime during these years, the finding evaporated – the declines from arresting so many immigrants were no greater than the general declining trends across the country, and the arrests and deportations thus had “no observable effect on the overall crime rate.”

Meanwhile, those federal agents focused on arresting immigrants aren’t working on other cases and stopping other crimes. According to Pew research, immigration-related offenses fueled half of all federal arrests in 2014. As those numbers rose, arrests for property crimes, gun crimes, drugs, and parole violations- all much more likely to result in violent recidivism – fell over the last decade. In fact, just five regions along the southern border accounted for 40% of federal arrests, which means that the rest of the country is getting underserved by federal law enforcement.  That’s the kind of skewed priorities that plays a role in violence growing unchecked in a handful of municipalities, such as Chicago.

Effective policing relies on communities feeling comfortable around police, because it is members of the community who know where crime is occurring and often, who is committing it. A literature review we conducted at the Carnegie Endowment found that problem-oriented policing, a technique that relies on community members to work with police on community-determined crime-control goals, was among the more effective techniques. The Campbell Systematic Reviews — nonpartisan reviews of contested policy issues with seed funding from the conservative-leaning Smith Richardson Foundation — agree that this form of policing is one of the only models with modest, but effective results.

Another policing technique that has gained ground since September 11 and has a growing success record in the United States draws from counter-insurgency theory. Under the title “Intelligence-Led Policing.” police, not community members, determine the focus of departments, but rely heavily on community members for information on who the violent criminals are, where they are based, and when they are going to strike. When the Bureau of Justice Assistance compiled case studies on intelligence-led policing to help fight endemic violence, often from gangs and drugs, in cities from Milwaukee to Palm Beach and Richmond to San Diego, the results were strikingly positive.

One of the only places the technique failed was in Phoenix – that’s right, the metropolis where Joe Arpaio was sheriff. The failure was partially due to money problems, but it may also have had something to do with the extraordinary climate of fear created by law enforcement led by a man like Arpaio. These successful techniques rely on communities feeling comfortable talking to police, including informants and “snitches.”  Such information, dialogue, and intelligence is highly unlikely to occur when communities fear deportation, having their status checked, or even just being treated rudely by the police due to the color of their skin.

In sharp contrast, making immigrants – including illegal immigrants – feel comfortable around law enforcement is likely to enable police to gain intelligence on the truly bad apples among them, as well as the criminals preying on them. In New York, people in neighborhoods with high concentrations of immigrants were both less cynical about law enforcement and more cooperative with police. It is difficult to prove causation, but those two factors often go hand-in-hand.

There is also ample evidence to show that when tough-on-immigration measures kick in, many immigrants become scared to turn to police for fear of harassment if they are legal residents, or deportation if they or members of their family are not. In Chicago, a 2012 survey of over 2,000 Latinos found that 45% were less likely to volunteer information about crimes to police, and 44% were less likely to contact police if they were victims of a crime, out of fear that police will ask them about their immigration status, or the status of family and friends. Among the undocumented community, 70% were unwilling to contact police if they had been victimized. Even 28% of Latinos born in the United States reported being unwilling to contact law enforcement out of fear they would ask about the immigration status of people they knew. That fear of law enforcement lets criminals and criminal syndicates stay on the streets to victimize others – immigrant and native-born alike.

Cutting DACA, Increasing Crime

If President Trump really wanted to fight crime and make America great again, his first step would be to increase immigration and legalize the status of those here illegally. President Ronald Reagan’s choice to do just those things with his 1986 amnesty probably had a hand in the vast decrease of crime in the following decade. A study of Reagan’s 1986 amnesty found that for every 1 percent of a county’s population who received legal status and came out of the shadows, violent and property crime rates fell.  Had Reagan not accompanied these immigration changes with counterproductive war on drugs policies, the improvements in violence might have happened on his watch rather than under President Clinton.

That study’s researchers attributed the fall in crime following amnesty to being able to work legally. That is, of course, the precise goal of the DACA program. Conversely, in San Antonio neighborhoods where many immigrants failed to gain legal status from the 1986 amnesty, economically-motivated felonies rose.  Sadly, it’s a fair bet that if crime rises following Trump’s decision to rescind DACA, he will blame it on the immigrants themselves, rather than his decision to drive Dreamers to despair and abolish legal avenues for people who have been living in the U.S. since childhood to earn a living.

The old canard that immigrants cause crime was first debunked in 1924 by Edwin Sutherland, the author of the criminal justice bible, Criminology.  But the idea keeps rising from the dead. It is time for reasoned lawmakers and fact-based policy experts to make a stand, and nothing could be more suitable ground than the DACA program. What happens with DACA could redefine America’s relationship with its immigrant population for years to come. The cost to America could be an end to our historically low crime rate, putting our country on the path to rising violence and making Trump’s overblown fears self-fulfilling.

This article was originally published in Just Security.