ICE Showdown: Who’s Really In Charge?

ICE badges on an American flag background — Photo: Mehaniq / Shutterstock

Campaigns turn complex intergovernmental questions into stories of control and culpability; the Blakeman–Hochul–Mamdani fight is a textbook case where rhetoric outruns proof on causation and subordination, even as real policy differences on immigration enforcement are unmistakable.

The Short Version

Blakeman’s “puppet” charge is documented rhetoric; the evidence shows political alignment between Hochul and Mamdani, not control.
The core policy divide is cooperation with ICE: Blakeman touts local partnerships; Hochul and Mamdani emphasize limiting “overreach.”
Claims that immigration policies caused a crime “epidemic” rely on anecdotes, not validated statewide trend data or causal analysis.
The broader research literature finds sanctuary-style limits do not increase crime, undercutting sweeping causal claims.

What is actually established, and what is not

The public record captures two distinct realities. First, the rhetoric: at a late-October rally, Bruce Blakeman labeled Governor Kathy Hochul “Mamdani’s puppet,” and on the debate stage he tied immigration to crime with vivid examples. Those statements are part of the campaign file and are not in dispute as quotations. Second, the policy posture: Hochul has described herself and New York City Mayor Zohran Mamdani as “great partners,” and the governor has appeared with Mamdani and the Attorney General to oppose aggressive federal immigration enforcement and reaffirm certain protections for New Yorkers. Those are on-the-record positions by the sitting administration. What is not established is subordination—no documentary trail shows Mamdani directing gubernatorial decisions—or a verified causal link between Hochul’s immigration stance and an “epidemic” of crime statewide. The available coverage supplies assertions and incidents, not the denominator, time-series, or identification strategy an analyst would require to show policy-driven crime effects.

That distinction matters. “Puppet” suggests control; “partnership” concedes alignment. The former is a charge of subservience; the latter is a standard intergovernmental relationship claim in a mayor–governor dynamic. Absent calendars, directives, or decision records showing control, the evidence supports cooperation, not capture.

The real policy divide: cooperation with ICE versus limiting “overreach”

New York has long navigated the boundary between local policing and federal immigration enforcement. Blakeman situates himself on the maximal-cooperation side, defending Nassau County’s relationship with ICE as a model and tying that cooperation to public safety. Hochul and Mamdani, by contrast, frame the state’s role as protecting New Yorkers’ rights against “overreach,” signaling limits on how far state and city agencies will go to assist civil immigration enforcement while still insisting the state is “not and never will be a sanctuary for violent criminals.” Those are meaningful, operational differences: they shape detainer responses, information-sharing, and whether local resources are used to effectuate federal civil priorities versus focusing on state criminal law.

The administration’s stance has taken specific form—public events opposing aggressive ICE tactics and support for legal defense for unaccompanied minors—while its spokespeople stress ongoing notification to immigration authorities when noncitizens are serving state prison sentences. Blakeman’s critique reframes those choices as permissive on crime; Hochul’s camp casts his position as licensing federal agents to “run rampant in our streets.” Each is a political translation of a technical coordination question: when, how, and to what extent should local actors help execute civil immigration law.

Crime causation: anecdotes are not analysis

High-salience crimes commit political gravity. They should not, however, be mistaken for causal inference. Linking state immigration posture to a surge in crime requires, at minimum, incident-level data over time, clear policy-change dates, credible measures of exposure, reliable identification of defendants’ legal status, and methods that can separate policy effects from confounding trends. The campaign materials on both sides do not supply that spine. They offer cases, accusations, and counteraccusations, but not a statewide analytic record that would survive basic methodological scrutiny. On the specific Marble Hill ICE-involved shooting that Blakeman cited as the preventable fruit of Hochul’s policy, the public dossier lacks operational records tying a gubernatorial choice to that confrontation. Counterfactual claims without the plan books and after-action reports are simply claims.

When researchers have tested the broader proposition—do sanctuary-style limits on cooperation increase crime?—the results are consistent: they do not find statistically meaningful increases in violent or property crime attributable to such policies, even as deportations, particularly of people without criminal convictions, decline. That literature does not adjudicate every New York incident, but it places a heavy evidentiary burden on sweeping assertions that state-level limits on ICE cooperation “cause” more crime. If a campaign wants to reverse that presumption, the path is empirical and transparent, not rhetorical.

How we got here: personalization as a strategy

Personalization is not a bug of modern politics; it is the operating system. Turning a coordination debate into a story about who controls whom—“Mamdani’s puppet”—does two things at once. It suggests motive and erodes the target’s agency. It also collapses a multi-layered policy space into a single antagonist, keeping the conversation on identity and allegiance rather than statutory architecture or data. Mamdani’s own responses, including calling Blakeman’s attacks bigotry and racism, complete the polarization loop, pushing the discussion further from falsifiable claims and toward moral verdicts about opponents. None of that resolves the central empirical questions a voter might care about: which procedures reduce victimization, which protect due process, and at what fiscal cost.

There is also a structural asymmetry at work. The documents that would clarify cause and effect—detainer logs, incident-level crime data merged to legal status, ICE operational plans—sit in agencies that do not release them readily. Campaigns fill the vacuum with message. Analysts, meanwhile, must live with fragmentary visibility until litigation, audits, or proactive transparency pry the files loose.

New: Governor @KathyHochul responds to criticism from her opponent @NassauExec who said she’s helping him by campaigning with Mayor @ZohranKMamdani

“It's kind of funny because Mayor Mamdani is more popular in the rest of New York State than Bruce Blakeman is” Hochul says pic.twitter.com/YRyyjyOE7J

— Morgan McKay (@morganfmckay) October 23, 2026

What evidence would actually settle the arguments

Three workstreams would transform this from a proxy war into a testable record. First, build a dated decision timeline mapping mayoral and gubernatorial statements against concrete directives, appropriations, and agency guidance; look for instances where Hochul diverged from Mamdani’s preferences to test the “puppet” thesis. Second, run a pre-registered, longitudinal analysis of statewide crime, arrest, and conviction data aligned to immigration-policy changes, with uncertainty estimates and clear treatment definitions; incorporate legally obtainable indicators of immigration status and distinguish allegations from adjudicated outcomes. Third, on sentinel events like the Marble Hill shooting, collect the operational plan, warrants, use-of-force reports, and body-worn camera footage to assess whether state-cooperation rules plausibly altered the chain of events. Those are ordinary tools of public administration research; they are simply underused in campaigns.

How to read the next round of claims

An informed reader can navigate this terrain with three filters. Separate rhetoric from record: quotations establish what was said, not whether it is true. Demand the denominator: single cases cannot carry system-level claims. And match the claim to the appropriate evidence: control demands directives; causation demands data and method. Applied here, those filters yield a sober conclusion. The partnership between Hochul and Mamdani is real and on the record; control is unproven. The policy divide over ICE cooperation is substantive and consequential; its public-safety effects, in aggregate, are not established by the campaign’s anecdotes and are inconsistent with the best-available national research. That is not a punt—it is the discipline that keeps policy from disappearing into slogans.