A central component of the protests that have followed the death of George Floyd has been a call for broad reforms to police departments. Floyd died after a Minneapolis police officer knelt on his neck for almost nine minutes, sparking new outcry over police practices and the interactions of law enforcement officials and black Americans.

One reform that has moved to the forefront of the debate is a demand that police departments be defunded — not necessarily by stripping all funding but at least by shifting responsibilities from departments and reducing government spending dedicated to them. The idea is that the role of police officers is far broader than simply enforcing laws and protecting the populace, introducing more occasions on which armed police interact with the public.

Intuitively, one might worry that reducing police spending would lead to a spike in crime. A review of spending on state and local police over the past 60 years, though, shows no correlation nationally between spending and crime rates.

In 1960, about $2 billion was spent by state and local governments on police. There were about 1,887 crimes per 100,000 Americans, including 161 violent crimes. By 1980, spending had increased to $14.6 billion — and crime rates had soared to 5,950 crimes per 100,000 Americans and 597 violent crimes. Over the next two decades, those rates thankfully fell, down to about 4,120 crimes per 100,000 people and 507 violent crimes. Spending spiked to more than $67 billion. Eighteen years later — by 2018, the most recent year for which full data are available — crime rates had fallen further to 2,580 crimes per 100,000, including 381 violent crimes.

Spending that year topped $137 billion.

The figures above aren’t adjusted for inflation. If we make that adjustment, the pattern since 1960 looks like the chart below: Crime and spending increasing at a similar pace until the early 1990s, when crime rates began to drop but spending kept soaring.

This is slightly an apples-to-oranges comparison because the crime rates are adjusted for population. If we control the spending for population, too, the pattern looks like this: climbing spending from 1980 to 2010, with crime rates falling from the early 1990s on.

Even in recent years, when national spending per person has dropped, crime hasn’t risen. In 2006, the United States spent $386 per person on state and local police, with a crime rate of about 3,800 per 100,000 people and a violent crime rate of 474 per 100,000. In 2010, spending rose to $412 per person and the crime rates were down to 3,350 per 100,000 overall and 405 violent crimes per 100,000. In 2012, spending was back down to $389 per person — but the crime rates had fallen further, to 3,256 per 100,000 overall and 388 per 100,000 overall.

If we look at how spending has changed relative to crime in each year since 1960, comparing spending in 2018 dollars per person to crime rates, we see that there is no correlation between the two. More spending in a year hasn’t significantly correlated to less crime or to more crime. For violent crime, in fact, the correlation between changes in crime rates and spending per person in 2018 dollars is almost zero.

If we assume a delayed effect — after all, after the drop in spending in 2012, violent crimes were up by 2016 — there’s still no clear link. Assuming a four-year delay, there’s still no significant correlation between changes in spending and overall or violent crime rates.

This is a very top-level assessment, admittedly. There are certainly places where an increased focus on and investment in policing has driven down crime rates. There are others, unquestionably, where more spending has not led to improved outcomes.

The challenge for elected officials, of course, is the perception. No one wants to be the mayor or city councilor who advocates shifting money away from the police right before something tragic happens. There may not be a national correlation, but that finding is battling a powerful opponent: political self-preservation.