Nearly three-quarters of marijuana processed for sale in Colorado evaded pesticide testing, and contaminated products sat on store shelves for as long as three years before regulators issued a public warning. Those are the headline findings of a state audit that portrays the Marijuana Enforcement Division as an agency systematically falling short of its core consumer-protection mandate.
A Testing Gap of Striking Proportions
According to ProPublica, the audit documents failures across nearly every stage of the MED’s oversight process. From 2023 through 2025, close to 75% of processed marijuana destined for sale bypassed pesticide testing entirely. During the same period, contaminated products that were the subject of an advisory remained on shelves for more than seven months on average before being addressed.
The timeline problems compound. In 2025, the agency took three and a half months on average to issue a public advisory once it had opened an investigation into potential contamination. Even after investigators internally determined an advisory was warranted, it took the agency another month, on average, to publish one. Contaminated products remained available to consumers throughout those delays.
Inspection failures deepened the exposure. The audit found regulators did not meet regular inspection schedules and failed to conduct sufficient random checks at dispensaries for contaminated products. The audit report was direct in its assessment: “These issues limit MED’s ability to meet its core objective to identify product safety risks and to help ensure that licensed marijuana businesses meet product safety requirements. As a result, there is an increased risk that marijuana products sold in Colorado may have contaminants or that the potency listed on the labels is inaccurate.”
Regulatory Ceilings and the Burden Left on Consumers
The audit’s most damaging finding may be structural rather than procedural. A survey of MED investigators found that more than half said there are, or may be, widespread marijuana product safety issues in Colorado’s regulated market. That figure, drawn from the agency’s own investigators, suggests the gaps are not incidental.
Zigmas Pekarskas, Co-Founder and CEO of Stake Hunters and a longtime observer of regulated risk markets, sees a familiar dynamic in the audit’s conclusions. The honor-code finding, he argues, exposes a ceiling that regulatory frameworks tend to reach regardless of intent.
“The audit makes plain that a regulator’s presence does not equal a regulator’s guarantee. Consumers in any regulated market eventually carry a residual verification burden that no agency can fully absorb.”
Pekarskas compares this with sports betting, where even bettors who follow StakeHunters daily betting tips need to check that they are using a properly licensed platform. As in other regulated industries, oversight has its limits, and regulators cannot guarantee every aspect of a user’s experience.
Colorado’s Honor-Code Testing System
The mechanism behind the testing gap is the self-sampling model. Colorado permits marijuana businesses to collect their own samples for safety testing, a practice MED investigators described to auditors as relying “largely on an honor code” to ensure product safety. More than half of states with regulated marijuana markets have moved beyond this model, instead requiring private labs or independent third parties to collect samples.
Colorado considered a similar overhaul, but abandoned it. MED deputy senior director Kyle Lambert told legislators the agency opted against requiring private lab personnel to collect samples, citing geographic challenges for rural businesses. “It was decided now wasn’t the right time to propose that as a policy change,” Lambert said.
The existing system has additional flex points. Colorado allows marijuana businesses to test far less frequently than required by state guidelines if they can demonstrate consistent compliance. Manufacturers can also claim further testing exemptions, including for certain marijuana-derived skin and body products, with limited MED oversight of whether they actually qualify.
State Sen. Byron Pelton, a Republican from Sterling, was blunt about the incentive problem this creates. “If I’m a shady character, I know which products I’m going to go sampling and which ones I’m not,” he said. “That’s a pretty gray area to me.”
MED’s Response and a Case Study in Delayed Action
MED senior director Dominique Mendiola said in a statement released the night the audit was published that the review “allows us to take a critical look at how we can continue to carry out our work, consistent with our commitment to integrity, safety, and continuous operational improvement.” The agency said it would assess testing exemptions currently allowed under state law.
That self-assessment sits in uncomfortable tension with one of the audit’s more concrete examples. It took 15 months for a manufacturer to discontinue a marijuana sleep aid pill after consumers warned regulators the product was harming their livers. The pills contained herbal extracts that researchers say can cause liver damage. The gap between the first consumer warnings and the eventual discontinuation illustrates precisely the kind of delay the audit documents across the wider system.
The auditors also found that voluntary recalls did not succeed at stopping some businesses from selling potentially unsafe product during the recall period. Whether MED’s promised reforms, including a reassessment of testing exemptions, will close gaps that its own investigators believe may be widespread remains an open question the audit leaves squarely on the table.
