Nearly 1.3 million people voted in 2012 to legalize recreational pot in Colorado, taking close to 55 percent of the vote and putting the state at the forefront of America’s marijuana revolution. By doing so, Coloradans effectively turned a shady, stigmatized practice—a hobby associated with airheads and slackers—into a public service that would be invested in something universal and of undeniable worth: school improvement.

The law, which went into effect last January, stipulates that the first $40 million raised from taxes on the sale of recreational pot help pay for construction to improve Colorado’s “crumbling” public-school buildings, a problem the state has long struggled to address because of limited funding. This is a familiar model. In other states, like California, schools are funded through another formerly taboo practice: lotteries.

In Colorado, the public-education benefit of legalization served as a major selling point for the proposal during election season. Pro-legalization advocates even aired ads with slogans like, “Jobs for our people. Money for our schools. Who could ask for more?” and  “Strict Regulation. Fund Education.” Politically, the tactic seems to have paid off given that, in the year or so that’s passed since the law went into effect, Coloradans continue to cite public funding as the greatest advantage of legalization. A 2014 survey found that most respondents saw an “upside” to legalization, with more than half of those adults identifying “tax revenue” as the greatest benefit (versus other “upsides” such as easier access, fewer arrests, and tourism). And more than half of all the survey participants said that public schools should receive the most funding from marijuana tax dollars; public schools were one of six options respondents could choose from.  

But whether legalizing recreational ganja will ever make a dent in Colorado’s school-construction backlog is far from certain—and the consequence isn’t only that schools won’t get the extra money they might have hoped for. Analyses of the policy’s fiscal impact, including a widely cited study by Colorado State University’s Colorado Futures Center, conclude that the excise tax earmarked for schools is hardly on target to generate that annual $40 million benchmark. The latest calculations estimate that funding this year will be about half that. One theory is that the tax makes recreational weed a lot more expensive than the medical stuff and might be incentivizing people to stay in the “gray” medical-marijuana market.

What's more, in an awkward (and perhaps embarrassing) twist, all that money could be lost. That’s because, under Colorado’s “Taxpayer Bill of Rights,” if in any given year the state reaps more tax money than revenue forecasters had projected, the state must return that extra revenue to taxpayers. This year, the provision will be triggered because—even though the pot money came in lower than expected—the state collected more tax revenue overall thanks to other industries such as energy and oil. Lawmakers are now crafting a bill that would ask voters this fall to approve an exemption to that provision for the pot tax.

Still, there are other caveats. As one official who helps oversee the state’s school-construction grant program told the Vail Daily last November, the policy’s education focus gives the false impression that schools are getting the funding they need: “Yes, we got some money. The last grant round, we got to fix some boilers and roofs, but legislators are still using that $40 million number. People believe that the facilities problem in Colorado have been fixed by this and it hasn’t.” Indeed, new research on the politicking around the legalization campaign offers further insight into the realities of the marijuana tax, suggesting that the use of education as a selling point for legalization could even undermine long-term efforts to improve Colorado’s public-school system.

The preliminary research, out of the University of North Carolina at Chapel Hill, offers the latest case study within a wider body of research on the exploitation of education “as a topic to sell voters on policy proposals that condone previously illegal behavior,” the analysts write. As past studies reveal, using education to justify the legalization of formerly condemned activities often serves as a selling point in name only.

A few years back, a Colorado school facilities assessment projected that construction needs on campuses across the state would amount to more than $17.8 billion by 2018. Under Colorado’s new law, revenues from the marijuana excise tax feed into an existing grant program that funds school construction and renovation projects—a program that has been funded primarily by land and lottery proceeds. (The UNC researchers didn’t focus their analysis on Colorado’s lottery, as the vast majority of the state’s lottery revenues go toward park conservation and wildlife.) During the 2014 fiscal year, the grant program’s revenue grew to more than $95 million from about $72 million, according to state Department of Education data. But the pot taxes only accounted for a sliver—$3 million—of that growth. Then, in 2014, the marijuana proceeds contributed just $13 million to the fund. Though more research is needed, the analysts write, the funding stipulated in the recreational-pot legislation is “miniscule” compared to school construction needs. And “it’s also important to remember that [school construction is] not the only need in funding in education,” said Brooke Midkiff, one of the UNC education-policy researchers. “A lot of funding needed in other areas as well.”

Curiously, as Midkiff and her co-author point out in the paper, 2014 revenues from the lottery fell by nearly half from the year before, an unusual drop whose timing largely corresponded with the legalization of recreational marijuana. Whether the two phenomena are at all related isn’t clear; a number of factors could explain the decline in participation, and there’s no evidence demonstrating that the introduction of a pot fund for schools somehow compelled gamblers to stop buying tickets. But the trends do resemble what researchers have observed in other states—researchers focusing, as it happens, on the legalization of lotteries in the name of public schools.

In a recent study titled “Education’s Gambling Problem,” for example, the University of South Carolina economist Daniel Jones explored the impact that lotteries introduced to support public schools have on voluntary donations to education. He found that donations fall once a lottery is introduced—a result “driven by [the] donors’ response to the new (highly publicized) government revenue source” rather than a drop in fundraising efforts.

Meanwhile, a 2007 study found that, of the 42 states that operated lotteries then, eight earmarked all the net proceeds for K-12 education, while another six included postsecondary education as a recipient. The researchers analyzed a sample of six of those states and found that lottery revenue corresponded with increased education spending in only one of them, though the reasons for that anomaly are difficult to assess.

Its open to question whether Colorado’s experience will see parallels elsewhere across the country. Washington legalized recreational pot the same year as Colorado with some revenue earmarked for education and the general fund, but because the law has taken longer to implement its fiscal impact has yet to be analyzed; Oregon and Alaska are rolling out their laws this year. In the meantime, legalization campaigns are underway in a number of other states, and many of them could seek to earmark any tax revenue for schools or other public services. Phyllis Resnik, the lead economist for the Colorado Futures Center, told the Pew Charitable Trusts’ Stateline last September that she hopes other states learn from Colorado when designing their laws: “All these promises that get made about free money to solve problems, marijuana is not going to be that.” In other words, while marijuana may be a solution to many problems, education funding probably isn’t one of them.