What a School District's Enrollment in ICE's 287(g) Program Reveals About the Messy Realities of Policy Diffusion

Too many unanswered questions in ICE's enrollment of Caney Valley School District in the 287(g) immigration enforcement program, highlighting ongoing issues with transparency and accountability.

Last week Andrew Thrasher shared an alarming finding from his scrupulous monitoring of ICE’s 287(g) program, which delegates immigration enforcement authority to local agencies: for the first time ever, a K-12 school district had signed a 287(g) agreement. The agreement was signed on March 13, 2026, and included in ICE’s online spreadsheet on Monday, March 16. Andrew announced this finding immediately given his understandable surprise. Even more surprising was the fact that the signed agreement was listed in ICE’s regularly-updated spreadsheet as a Task Force Model agreement, the most aggressive and expansive type of 287(g), which would mean that the Caney Valley school system’s newly-created police department would be able to take enforcement actions against parents and children in the district. Totally unprecedented.

Yet as quickly as Caney Valley’s 287(g) appeared, it disappeared. In the Wednesday 287(g) spreadsheet update, Caney Valley was gone again. I asked Andrew to discuss this with me on Substack Live last Wednesday because I hoped that reporters would take a closer look at what happened. And they did. I am grateful to the anonymous regional contacts and reporters that provided Andrew and me with additional context, and to the prolific Greg Sargent for reaching out to the school district. Andrew’s work prompted two news articles, including one published by the local NPR station yesterday, a slew of background conversation, email threads and text messages, and an update from Greg on Bluesky. I encourage you to read Andrew’s thorough and excellent write-up yesterday, that comes with all of the receipts, titled "Un-Caney Valley” (chef’s kiss title, my dude).

Long story short, the school district does not currently have an active 287(g) agreement and is not interested in pursuing one further. That’s good news. But the story is not over. Not even close. There is much left to untangle and, unfortunately, the reporting on this story has only added, rather than alleviated, the confusion. This essay parses details about this case that others have missed and draws larger lessons from this unique sliver of a window into the messy everyday realities of 287(g) policy diffusion.

Let’s start with the conflicting statements about the basic question at the heart of this story: did a school district actually sign a 287(g) agreement. There are two key people to keep straight in the local district: District Superintendent Steven Cantrell and school district Police Chief Michael Coates.

When asked about the 287(g) agreement by Ben Fenwick from Oklahoma Watch, Cantrell said: “He [Coates] requested official information but did not sign anything.” Cantrell also told Greg Sargent, “Our district has not, nor will not enter into an agreement with ICE.” Cantrell, also told NPR that, “the district never entered into the agreement because he [Cantrell] is the only one with the authority to sign binding contracts on the district's behalf.”

But that’s not what Chief Coates said. In an email that was not published, Coates said, “While we initially considered the agreement as a potential training opportunity, we ultimately decided to terminate it once we determined it was not applicable to our role.” This is reflected in NPR’s reporting—but only obliquely. The NPR article says that the school district “seemingly sign[ed] an agreement to deputize its police department as immigration enforcement officers, and then back[ed] out of the partnership days later.”

And Cantrell's quotes above contradict his own other statements given to the press. The same NPR article, confusingly titled “Oklahoma school district hoped for training, almost agreed to a partnership with ICE instead,” goes on to say that “district Police Chief Michael Coates mistakenly signed an agreement deputizing them with federal immigration enforcement authority through U.S. Immigration and Customs Enforcement (ICE), according to the district superintendent Steven Cantrell."

Which is it? Did the agency sign a 287(g) agreement, then back out? Or did they never sign an agreement in the first place, and ICE lied publicly about their participation? The confusing statements by Coates and Cantrell, and internally-contradictory NPR reporting on this has not actually provided any answers. Understanding whether a senior law enforcement official signed a piece of paper and submitted it to the federal government is a factual matter and one that should be cleared up in a single comprehensible statement, preferably with evidence of whatever he did send to ICE. That still hasn’t happened.

But wait, there’s more. Let’s look at the description of how Mr. Coates learned about the 287(g) program in the first place. Mr. Cantrell told Oklahoma Watch and NPR that Chief Coates first learned about the 287(g) agreement during a sexual assault training.

“Cantrell said the training event attended by Coates was to train officers to prevent sexual assault. There, Coates was contacted by ICE about training, Cantrell said. Cantrell said Coates had recently been confronted with adults on the campus after normal hours who did not have identification and could not speak English. He said Coates wanted more information on what to do in such a situation.“ —Oklahoma Watch

“During a training on how to approach sexual assault investigations earlier this month, the district Police Chief Michael Coates mistakenly signed an agreement deputizing them with federal immigration enforcement authority through U.S. Immigration and Customs Enforcement (ICE), according to the district superintendent Steven Cantrell.” —NPR

When I read this, I was baffled. Why would ICE be at a sexual assault training for cops? Are they so overstaffed that they are sending officers to routine trainings to talk about 287(g)? Taxpayer-funded police trainings typically have mandatory requirements related to hours and types of training. If ICE is coming to a local training to talk about an entirely separate matter, that time would have to be taken out of mandatory hours and probably should not be (or would not be allowed to be) supported with taxpayer funds. It was entirely plausible given the political moment, of course, but something didn’t add up.

That’s when I went to the Facebook page for the event that shows officers at the sexual assault response training hosted by the Oklahoma State Bureau of Investigation at a nearby university’s regional campus. The Facebook page shows that the sexual assault training was on March 13, the same date that ICE says the 287(g) agreement was signed. The photos appear to show officers from a variety of agencies, given the diversity of badges in the room; it wasn’t Caney Valley alone. This is important, because whatever confluence of conditions led to Caney Valley signing up, it means that a lot of other agencies may have learned about 287(g) the same way and also might have been encouraged to sign up. To put it in more generalizable terms: this situation provides us with some unique insight into the mode of 287(g) policy diffusion which is bigger than Caney Valley alone.

One thing that has gone unnoticed so far is the fact that the host for the training, the Oklahoma Bureau of Investigation, participates in the 287(g) program. In fact, they were among the first wave of agencies to sign onto the Task Force Model after Trump took office. Andrew’s interactive 287(g) data tool shows that several state agencies in the state signed on less than a month after inauguration.

Currently 40 agencies in Oklahoma have an active 287(g) agreement.

This is important because the reporting above says that Mr. Coates was contacted by ICE (or, let’s say more broadly, had some kind of interaction with ICE) at the sexual assault training. Oklahoma Watch wrote:

“District Superintendent Steven Cantrell said his school’s police chief was misled by ICE during a recent training event, during which he asked for more information but was instead sent a 287(g) agreement and asked to sign it. … [At the training], Coates was contacted by ICE about training, Cantrell said. Cantrell said Coates had recently been confronted with adults on the campus after normal hours who did not have identification and could not speak English. He said Coates wanted more information on what to do in such a situation.”

We can only make sense of the information we have, but I think it’s important to understand whether ICE is personally going out to routine officer trainings to recruit participation or whether this was driven by other state or local 287(g) agencies engaging in peer recruitment. Cantrell indicated that Coates was “misled” about the 287(g) program—but by whom?

ICE could have been present at this training. The training was held closer to Tulsa than Caney Valley and the latest data on ICE field offices made possible by the great work of Jacqueline Stevens and Nicolette Glazer do show that Tulsa has a field office with 287(g)-related responsibilities.

As a matter of geographic convenience, this makes sense—but it doesn’t make sense given the totality of the circumstances. The 287(g) form is available online and submitted to a public email address. It’s not as if ICE would have to physically deliver any paperwork and there’s no indication that ICE needed to come in person to explain things to Coates; Coates’s clearly left the training that day without a good understanding of the 287(g) program. It’s possible that an ICE officer drove out just to say, “sign here” with no explanation or they had some kind of phone conversation or someone at ICE convinced Coates to sign and submit the form electronically.

However, the fact that there were other 287(g) agencies in the room makes it plausible that the seminar’s trainers or someone in attendance who already had a 287(g) agreement suggested the program to Mr. Coates. In this scenario, it may not have been ICE directly, but a kind of proxy representative, a colleague at the training who said, “worried about non-English speakers with no ID? Fill out this form online and send it into ICE, they’ll hook you up.” It’s also plausible (though again speculative) that rather than say this to reporters, who would likely want to go on and talk to Coates’s other colleagues, too, it was easier to simply gesture toward ICE. I’m saying this as someone who was a police officer in the Navy and has some sense of what it’s like inside these types of institutions. The last thing you would want to do if you publicly fumbled a serous piece of paperwork is make it look like you’re blaming senior colleagues at an agency you need to collaborate with professionally and probably get beers with after hours.

We don’t know exactly who turned Caney Valley on to the 287(g) program or who exactly misled Mr. Coates. It is also entirely possible, as some of the reporting suggests, that Coates was only asked to sign the document and that ICE fabricated the entry in the 287(g) spreadsheet with no signed agreement. The conflicting statements around whether a form was signed have been enormously unhelpful in pinpointing what happened.

To Coates’s and Cantrell’s credit, once they learned about the 287(g) agreement, they ran in the opposite direction immediately. Cantrell told reporters that he had no interest whatsoever in pursuing immigration enforcement powers of any kind and Coates used a more neutral “not a good fit” approach to distancing his agency from the 287(g) program.

“Our mission statement, the very first sentence, says that we're going to provide a safe and positive learning environment for all of our kids,” Cantrell said. “So I don't care if they're Hispanic, African American, Native American, they're going to get the same support that we can give anybody.” —NPR

“There’s never going to be a time where we’re going to go into a contract with ICE or any other government agency, other than something educational,” Cantrell said. “ The last thing that I want is for my parents to think that we’re going to be opening our doors to ICE and submitting information to them, or anything else.” —Oklahoma Watch

Coates told NPR:

“As a campus police agency, we do not operate a detention facility and are not engaged in patrol operations related to immigration enforcement,” Coates wrote. “Therefore, the broader scope and purpose of the agreement do not align with our mission or daily responsibilities.”

Cantrell more than Coates seems to truly understand the damage to students and parents if a school district were to sign up for immigration enforcement authority. Neither of these statements really addresses the question of how the brand new police department will handle interactions with non-English speakers or whether Cantrell and Coates are actually on the same page. But it does signal an admirable readiness to take in new information and adjust accordingly, which is what you want in leadership more than perfection.

At this point, you would be justified to ask: why focus on this seemingly insignificant detail when, as of now, everyone agrees that the school district isn’t participating in 287(g)? There are three parts to this answer. The first two concern the law enforcement agencies in question, both of which should alert the public and lawmakers to the glaring lack of oversight of one of the most controversial immigration enforcement policies on the books today. The third is more generalizable about how we understand the often opaque practices surrounding 287(g) policy diffusion.

Local transparency and accountability. If the chief of a police department signed a memorandum of agreement—which, according to him, it seems he did, although regrettably—without understanding what it meant, it represents precisely the kind of dangerous inattention that should raise questions about qualifications for leadership in a law enforcement agency, especially at a school. A 287(g) agreement does not hide its purpose: it’s right there on page one. An MOA should have at least passed across the desk of the school district’s attorney. If I were the superintendent, school board member, or voting member of the public, I would be inclined to dig further into the universe of possible agreements that Mr. Coates has unwittingly entered into to make sure this isn’t a pattern. I would further encourage a local resident to submit an Oklahoma public records request for copies of all communications between Mr. Coates and ICE and other federal and state agencies to discover what exactly happened here.

ICE’s lack of 287(g) oversight. If ICE signed and approved this 287(g) agreement with a school district—which the agency affirmed in public records that they did—it sends an clear signal that the agency is not only signing these MOAs with no review whatsoever, but that the agency may actually be eager to get immigration enforcement authority and ICE officers into public schools. I realize that more or less all formal oversight and accountability mechanisms at ICE have been abolished and that the sensitive location memo has been rescinded. But I also believe that most Americans would draw the line at having local resource officers wandering the halls asking kids where they’re from or yanking parents out of their cars in the pick-up line. On the off chance that ICE would like to claim that including this school district in its official list was a mistake, then the agency would be admitting that it, too, is inattentive to its own work and unable to properly validate and screen agencies on its list.

Credit to Andrew Thrasher.

287(g) policy diffusion and the need for legislative reforms. When the section 287(g) was added to the Immigration and Nationality Act, it created a mechanism to circumvent longstanding legal precedent that state and local law enforcement agencies shouldn’t be enforcing federal immigration laws. But in doing so, it added at least a minimal accountability provision that delegation of immigration enforcement authority needed to be explicit and in writing. The problem is, many law enforcement agencies I’ve talked to over the years have no idea what they are signing or what a 287(g) agreement actually means—someone just told them they should do it, and they did it. This is actually the crux of the question and the answer—whatever it is—says a lot about the often casual, handshake relationships that federal, state, and local agencies have with each other that often goes unnoticed by the public. In this case, we get a sliver of a window into kind of institutional inertia, acquiescence, manipulation, and groupthink that circumvents transparency and undermines trust. The lack of adequate oversight, which the Trump administration has no interest in, only further illustrates the dangers of this program. If a local police chief can accidentally sign a 287(g) and ICE rubber stamps it, it is evidence that, fundamentally, Congress’s safeguards on this program are wholly inadequate and the 287(g) program either needs significant revision or be done away with entirely.

This was just one example of weirdness around 287(g). There are 1,552 total 287(g) agreements in existence across 1,315 agencies. What would happen if we looked into every single one of them as closely as this one? How many other agencies have signed up for 287(g) not really knowing what they signed up for? How close are we to seeing a K-12 school legitimately sign up for a 287(g) with the intention of asking parents and students about immigration status? There are so many unanswered questions, but this is why we need to continue to pay attention to this policy and document its rapid and reckless expansion.

I want to close by saying that none of my own understanding of this or the reporting on this topic would have been possible without Andrew Thrasher’s excellent work on 287(g) over on Maxwell Commons.