Colorado Gov. John Hickenlooper has two facts in front of him: Since 2014 crime has been rising in his state, outstripping the national trend, and since 2014 recreational use of marijuana has been legal.
Whether the two are connected is hotly debated -- and if they are, then what? For the first time publicly, Hickenlooper told CNN he doesn't rule out recriminalizing recreational marijuana, even if that's a long shot.
"Trust me, if the data was coming back and we saw spikes in violent crime, we saw spikes in overall crime, there would be a lot of people looking for that bottle and figuring out how we get the genie back in," he said. "It doesn't seem likely to me, but I'm not ruling it out."
Data is now coming back. In 2016, the state's crime rate was up 5% compared with 2013, while the national trend was downward. Violent crime went up 12.5% in the same time while the national increase was less than 5%. But Hickenlooper isn't yet ready to pin the blame on the legalization of weed -- a step he opposed but has since embraced as the choice of his constituents.
"This is one of the great social experiments of the last 100 years. We have to all keep an open mind," he said.
Denver, the state's capital and largest city, is home to the lion's share of Colorado's recreational marijuana dispensaries. It has more than 170 of them -- more than the number of Starbucks, McDonald's and 7-Eleven stores combined.
Since 2013, Denver has seen its crime spike, too; the 2016 crime rate increased 4%, with violent crime up 9%.
The Denver Police say the data is inconclusive.
"[Property crime is] the biggest driver of our [overall] crime, and of our increases. So, can you attribute that to marijuana? I don't think you can," said Denver Police Commander James Henning. "The data isn't there."
The force has added more officers to police the illicit weed market that Henning says continues to grow.
But, Henning said, there is plenty of gray area when it comes to cataloging crimes that may or may not have a nexus to marijuana -- legal or otherwise.
"If a marijuana dispensary is burglarized, is that because it was a marijuana dispensary or ... if it were a liquor store or a stereo store would it have been burglarized as well?" he said. "The data is so tough to nail down and say this crime happened because of marijuana. It's just almost impossible to do that."
Two years ago, Denver Mayor Michael Hancock blamed legal marijuana for drawing people to a pedestrian mall downtown where violent incidents were happening. In one case, a transient swung a PVC pipe at people nearby. Police did not classify that crime as "marijuana related."
In Fort Collins, Larimer County Sheriff Justin Smith is one of the few law enforcement leaders in the state to publicly blame legal marijuana for rising crime.
He doesn't claim that smoking a joint makes you more likely to rob a bank. The connection between cannabis and crime is often indirect -- and not captured by official statistics, he said.
"It's not a causal thing," he said, arguing instead that legal weed is attracting a growing seasonal transient population -- a population that he said is more likely to commit crime. "Every third inmate in the [Larimer County] jail is a transient and you go by and ask them, and they'll tell you, we came here because of marijuana."
Smith -- a longtime opponent of legal weed who once led a lawsuit against Colorado's legalization -- also said the theory that legalization would end the black market in marijuana has not been borne out.
"That was one of the big promises [of legalizing marijuana] that if you regulated it, you would get rid of the problems that had traditionally been there with the illegal grows, but it's been really the opposite," he said.
Mason Tvert, a well-known pro-marijuana activist, sees things very differently, arguing it's irresponsible to even suggest there's a connection between rising crime and marijuana without hard evidence to prove the link.
"The only story here is that the evidence does not show marijuana or marijuana legalization are to blame for this increase in crime," he said.
Did marijuana bring a killer to town?
Smith's frustration reached a boiling point last summer when the body of 23-year-old Helena Hoffmann was pulled out of a lake in Fort Collins. Police said she had been raped and murdered walking home from an overnight shift at a nearby McDonald's. The man convicted in the case, Jeffrey Etheridge, is just the kind of person Smith is warning against.
Etheridge is a registered sex offender from Kentucky. From jail, he told CNN that he moved to Colorado in 2017 with his then-girlfriend because her brother worked at a marijuana dispensary. At the time of his arrest he was a transient, living out of his car in the park where Hoffmann's body was found. Etheridge pleaded guilty but now says he is innocent.
Hoffmann left behind a then-4-year-old daughter named Mary, now being raised by her father, Zach Denton.
"I remember Mary looking at us, and she goes 'did my mom die?' and that's really when it set in," Denton said about the day he broke the news of Hoffmann's death to his daughter.
He said Hoffmann would not want all homeless or transient people blamed for problems caused by just a few.
Denton thinks the bigger issue is that Etheridge, an out-of-state sex offender, was able to register his overnight address as Fort Collins' City Park, a place that is supposed to close at 11 p.m. and attracts children who come to play and swim.
From a bench in City Park built in remembrance of Hoffmann, Denton said there's a lot that could change: sex offender laws, transient laws, or even the rules on park access at night. Whatever does change, he said, will be Hoffmann's legacy.
In downtown Denver, the large -- and growing -- homeless population often gathers in Civic Center Park, next to the Statehouse.
The state's rate of homelessness rose 5.3% from 2013 to 2017, according to data from the US Department of Housing and Urban Development. Nationally, the rate of homelessness dropped 8.6% in the same period.
Tom Luehrs, the executive director of Denver's St. Francis Center homeless shelter, said he sees many people who came to Colorado hoping to work in the legal marijuana industry, only to find out it's not that easy. But he said there is another, smaller, group of seasonal transient people who seem to prefer life on the streets to an apartment and a job. While their presence predates marijuana legalization, it has increased since it became legal, he said.
"A lot of the people that we work with are wanting to get jobs, wanting to get housing, wanting to move out of homelessness, so then you have this other group that's kind of even belligerent and certainly not engaging and sometimes just very disrespectful. They don't care," Luehrs said.
Hickenlooper is skeptical that legal weed is to blame for increasing the homeless population.
"We're trying to get data on it. That's a difficult one to measure," he said.
The need for real data
The lack of solid evidence one way or another weighs on Hickenlooper, who can point to other things that have changed since legalization on January 1, 2014 -- like an economic hot streak -- without being able to say exactly what impact that has had.
"When you have that kind of [economic] growth, you attract all kinds of people and a lot of them are unsavory. Do they come for the marijuana? Or do they come because there are so many young people coming, there's a lot of money in the community and this is a great place to try and rob somebody? Again, more data. More data is the only way we're going to figure this out," he said.
That was his advice to California lawmakers last year ahead of that state's legalization of recreational marijuana.
"Spend the money to get a good baseline so that you can help guide the discussions and the real facts around this huge transformational shift in the way we address marijuana," he told CNN, explaining his message to lawmakers.
Case in point: Colorado's traffic fatalities where the driver tested positive for the active form of THC known as Delta 9 more than quadrupled from 18 in 2013 to 77 in 2016, according to the Colorado Department of Transportation. But those numbers are likely very misleading, because, according to Hickenlooper, the state didn't often test for marijuana in fatal crashes prior to legalization.
"That's not real data," he said. "We didn't use to measure it and now we're trying to measure something, so of course we see a lot more."
If and when the data does come in -- from Colorado, from California and elsewhere -- it will be studied intensively. And if the haze clears and there are strong signals that state legalization has hurt the community, Hickenlooper said Colorado's legal marijuana experiment may have to end.