An argument I included in Monday evening’s post, about the Colombian Defense Minister’s security performance, raised a few hackles on social media.

That post cited President Iván Duque’s crediting Defense Minister Guillermo Botero for a 2 percent reduction in homicides so far in 2019, compared to the same period in 2018. To refute it, I cited the work of Colombia’s Peace and Reconciliation Foundation, whose big annual analysis of the conflict, published in August, looked at the data for the first half of 2019.

Much of the 2018-19 decrease in killings, that report concluded, owes to shifts in the landscape of armed groups and organized crime—not to the Duque government’s security policies. In Colombia’s historically conflictive regions, homicides rose in 2018—then shrank in 2019 so far, as armed groups achieved regional monopolies, entered into non-aggression pacts, or otherwise altered their behavior, making many territories less violently disputed.

Some on social media found this argument insulting. A few analysts, some of whom have carried out security policymaking for the Bogotá municipal government, called me out for failing to credit to the work of Colombia’s security sector, especially those in charge of security in the main cities, which have seen important drops in homicides. For them, this argument crediting organized crime realignments lacks “responsibility,” was “politicized,” and “doesn’t stand up to the most minimal empirical analysis.”

I’m sorry, but I need to push back a bit. Let’s unpack this further.

First, I’m not talking about the progress made since 2010, or since 2002. I’m talking about the 2018-19 timeframe, for which Colombia’s government claims a 2 percent reduction in homicides year-to-date, after a 6 percent increase from 2017 to 2018.

Between 2002 and 2008 or so, and again between 2013 and 2017, Colombia achieved some very important decreases in homicides, unlike anything seen elsewhere lately in Latin America. And I absolutely agree that much of it owed to government policies, especially in the larger cities.

But then there was a hiccup in 2018. After the FARC’s demobilization and exit from many areas, homicides increased last year amid a violent reordering of organized crime and armed-group activity and a continued absence of state presence.

Daniel Mejía of the University of the Andes, a former Bogotá municipal security secretary, tweeted violent crime data charts appearing to show an inflection point after August 2018, when President Iván Duque was inaugurated. Mejia sees this as evidence of the new Colombian government’s actions.

Why would that happen right after August 2018? Did Duque and Minister Botero offer a superior recipe for dealing with insecurity? Was Colombia in need of a conservative government’s more iron-fisted approach? Or have lots of other, parallel, things happened in Colombia since August 2018?

These 15 months saw, for instance, a non-aggression pact form between ex-FARC and other groups in the violent port city of Tumaco, and in surrounding Pacific coastal areas used heavily for cocaine trafficking. It saw the ELN all but vanquish the EPL guerrilla/criminal group in the Catatumbo region. It saw a group called “La Mafia” consolidate its presence, avoiding aggression with FARC dissidents, in Putumayo. In Arauca last month, I was told the ELN and FARC dissidents had entered into a non-aggression pact. I heard the same in Chocó about similar arrangements between the ELN and FARC dissidents and, in some parts of the department, between the ELN and the Gulf Clan neo-paramilitaries.

These are just a few examples. A proliferation of these “new equilibria,” most of which date to after August 2018, can’t be dismissed as a potential reason for this period’s drop in homicides.

Second, I’m hardly talking about Colombia’s main cities, most of which for years have had homicide rates well below those of Baltimore, Detroit, St. Louis, and several other U.S. cities. There, talented security technocrats and honest cops have achieved strong gains.

In 2018, homicides continued to drop in cities like Bogotá, Medellín, Cali, and Barranquilla. But the country as a whole saw an increase. Something was seriously wrong in some smaller cities and rural areas where the drug trade and armed groups have thrived. But not the principal cities.

“In effect,” concludes the cited Peace and Reconciliation Foundation report, “the 2018 increase in violence took place in municipalities with the greatest institutional weakness, the presence of illegal markets, and low institutional capacity—that is, in rural areas that have historically been affected by violence. In the big cities and in coca-free municipalities, the violent homicide trend continues to diminish.”

But then in 2019, the trend reversed again: homicides are down nationwide. In its August report, the Peace and Reconciliation Foundation crunches the national medical examiner’s data from January through June 2019, comparing it with the same period in 2018. It finds 497 fewer homicides across Colombia during that 6-month period, which in fact is more than a 2 percent decrease.

But it gets more interesting. The Foundation identifies 281 municipalities (counties; Colombia has 1,100) that are “priorities for the post-conflict”: places where government presence is scarce and there’s a history of armed-group activity. Those municipalities made up 30 percent of homicides in 2018. Colombia’s five largest cities made up 28 percent, and the rest of the country shared the other 42 percent.

During the first half of 2019, the medical examiner’s data show these 281 municipalities with a surprising 13 percent fewer homicides: 243 fewer people were killed here compared to the first half of 2018. As noted, nationwide over the same period, homicides declined by 497, so almost half of Colombia’s homicide reductions during January-June 2019 happened in the 281 most troubled municipalities. Colombia’s other 800-plus municipalities, including the major cities, shared the other half.

That is a remarkable result. What’s the miracle in these historically abandoned corners of the country? It’s not a big increase in government presence: the PDETs, “Zonas Futuro” and other post-conflict plans to introduce government into these territories are still just getting off the drawing board. It’s not the genius of urban security planners, whose writ hardly extends to the “priority for post-conflict” municipalities.

That’s where the Peace and Reconciliation Foundation’s hypothesis comes in. Something has happened, in many long-abandoned regions, among the armed groups themselves, without regard to government policy. This makes sense to me, as the most plausible explanation for why homicides have declined in conflictive rural areas. I’m actually surprised, when referring to what’s happening outside of the cities, that it would even be that controversial. Meanwhile, the contrary evidence presented to me doesn’t knock it down.

(A bigger project for another day—I’m traveling right now—would be to use the Colombian National Police’s homicide statistics, downloadable here as big Excel spreadsheets, code the “priority for post-conflict” municipalities, and view the year-to-date data to see whether it differs from Peace and Reconciliation’s number-crunching, and if so how.)