The same architecture that produces an 80% reduction in California-style reimbursed skimming fraud also produces the conditions for measured fraud to migrate to channels the metric cannot reach — bot-driven card-not-present attacks, terminal-side bypass, and a builder layer that the criminal networks have every incentive to keep hidden from enforcement. Ohio’s planned chip-card transition, announced by Matt Damschroder, who directs Ohio’s SNAP program, as a $20 million, three-year fix “largely borne by the state,” will likely show a comparable curve. Whether that curve represents an outcome depends on which fraud channel the analyst is watching.

1. The pre-mortem of Ohio’s chip-card plan, written as though it has just been completed.

The Ohio Department of Job and Family Services has spent the better part of three years replacing magnetic-stripe EBT cards with chip-enabled cards across the state’s SNAP population. The transition cost at least $20 million, the bulk of it state-funded. Monthly reimbursed skimming fraud has fallen along the trajectory California recorded — from a peak figure into the low single-digit millions.

The decision point that broke was the assumption that terminal-side compliance would follow card-side compliance. The plan shipped chip cards. It did not ship an enforcement mechanism requiring merchants to disable swipe acceptance, did not mandate chip-on-failure fallback behavior, and did not address the 40% swipe share that California’s social-services department disclosed to The Wall Street Journal. Within 18 months of full rollout, investigative accounting would show a familiar pattern: skimmer devices recovered from Ohio terminals have been physically modified to block the chip slot while leaving the magnetic-stripe path intact, forcing users back to the swipe channel the migration was designed to retire.

The leading indicator that would have shown the breakage as it began: a divergence between reimbursed skimming fraud (falling) and total SNAP account fraud reported to the USDA Office of Inspector General (rising or flat), accompanied by recoveries of Bluetooth-enabled skimmer devices that were not present in the pre-migration baseline. Vincent Porter, a Secret Service financial analyst cited by the Journal, noted that “Devices can sit undetected on a payment terminal for a year” and that “a single skimmer can compromise thousands of cards.”

A second-order failure compounds the measurement problem. Federal reimbursement authority was withdrawn by Congress in late 2024, after the Agriculture Department said it had replaced $360 million in benefits diverted from 754,000 households between October 2022 and December 2024. With that instrument gone, Ohio’s plan inherits a measurement signal that no longer exists. Agriculture Department Inspector General John Walk told the Journal: “That was just from folks who actually called in.” He added: “Without reimbursements, ‘they’re probably underreporting even more than just a few years back.’” A state plan whose success criterion is “reimbursed fraud falls” is therefore evaluating itself against a vanishing baseline.

2. Probabilistic forecast: how much skimming fraud does Ohio’s plan actually displace?

Reference class: California’s reported trajectory, substrate-grounded. Monthly reimbursed losses in California fell from “nearly $21 million” in January 2024 to “$4.3 million” in July 2026 — an arithmetic reduction of roughly 79.5% over roughly 30 months.

A second reference class — the U.S. EMV migration for credit and debit cards following the 2015 liability shift — is broadly analogous. The substrate does not independently establish this analogue; the 2015 date and the post-migration counterfeit-fraud reductions associated with the EMV transition are widely discussed in industry literature but lie outside the supplied article. With that caveat, California’s arithmetic reduction of ~79.5% sits within the order of magnitude commonly attributed to that transition in industry-published analyses, supporting the comparability of the two migrations for forecasting purposes.

Inside-view adjustment upward (more displacement of physical skimming specifically): Ohio layered card-locking features in May — the article does not specify the year — prior to chip deployment, a capability the credit-card reference class did not have at migration. This adds perhaps 5-10 percentage points of incremental reduction in skimming incidents.

Inside-view adjustment downward (less net displacement of welfare fraud as a whole, more substitution): the credit-card reference class assumed a unified payment-ecosystem migration in which both cards and terminals were upgraded in step; Ohio’s plan is unilateral on the card side and depends on merchant terminals that, as California’s data show, retain a 40% swipe share. Mark Haskins, the former USDA special investigations branch chief, told the Journal: “Chip cards are not going to solve the problem.” Sabrina Feve, the San Diego-based assistant U.S. attorney, framed the residual exposure in technical terms: “EBT fraud will continue to be a vulnerability as long as benefit cards can be used with only the information stored on the magnetic stripe, without the added security provided by chip technology.”

Combined probability estimate: there is approximately a 65-75% chance Ohio achieves a ≥75% reduction in physical-skimming fraud as measured by state-reported incidents over the three-year rollout. There is a meaningfully lower probability — roughly 20-35% — that total fraud against Ohio SNAP households (skimming plus bot-driven account-number guessing plus cloned-terminal pass-throughs, all tactics the Journal reports as rising) declines by a comparable magnitude. The widening gap between these two probabilities is the operational risk: measured skimming fraud may fall while total fraud against Ohio SNAP households rises or holds.

3. Wicked-problem structure: why each layer of intervention produces its own adaptation.

The substrate supports five competing framings of the underlying problem, and each framing generates a different prescription.

  • Technology-framing. EBT cards rely on magnetic-stripe technology that the article dates to the 1960s. Haywood Talcove, chief executive of LexisNexis Risk Solutions’ government business, characterized the continued reliance in stark terms: “The insanity of letting people continue to use a glorified hotel-room key is akin to donating money to the Romanians.” This framing produces chip-card migration as the prescribed intervention.
  • Criminal-enterprise-framing. A loose hierarchy of Romanian organized crime groups, with builder-skilled engineers at the protected core, runs the operation. Maria Rîmniceanu, a prosecutor with Romania’s Directorate for Investigating Organized Crime and Terrorism, told the Journal: “We are going against them.” The Transylvania seizure — a 2,500-square-foot home taken from a man convicted under Romanian money-laundering laws — illustrates the leverage point: downstream proceeds, not upstream capability. Sabrina Feve’s framing of the criminal structure is also relevant: skilled skimmer-builders, she said, are “especially protected by the groups,” rarely using stolen cards lest they get caught.
  • Welfare-system-framing. With federal reimbursement lapsed, defrauded SNAP recipients “often must turn to family or food pantries until their next monthly allotment.” Damschroder’s framing: “The problem is large enough that it’s having a significant negative impact on our most vulnerable citizens.” Tristan Rader, the Ohio Democratic state representative who sponsored the chip-card legislation, added a moral characterization: “There’s a special place in hell for people who steal from the poor.” This framing produces victim-recovery policy as the prescribed intervention — which the data show has effectively lapsed.
  • Architectural-framing. Haskins and other fraud experts argue the EBT card itself is the wrong instrument and that mobile wallet payments would bypass the terminal attack surface entirely. This framing produces a longer-horizon redesign, with no operational timeline in the substrate.
  • Enforcement-priority-framing. The Journal reports that the White House’s fraud crackdown “has focused most visibly on Minnesota’s Somali community, with dozens of convictions dating to the Biden administration, though in recent years Romanian organized crime groups have been at the forefront of widespread benefits theft involving skimmers across the U.S.” The reimbursed-loss data the Journal cites — $360 million in federally replaced benefits over 26 months, alongside 70 Secret Service sweeps yielding 694 skimmers and an estimated $720 million in prevented fraud — and Michael Peck, head of the U.S. Secret Service’s Global Investigative Operations Center, characterizing the operational picture as “Sadly, business is booming,” together describe an enforcement focus that does not appear to track the locus the reimbursed-loss data identify. This framing produces a debate about resource allocation, but offers no operational lever against either skimming or account-number guessing.

None of these framings, taken alone, resolves the others. The technology-framing intervention (chip cards) shifts attack volume into card-not-present channels that the welfare-system-framing intervention (reimbursement) no longer measures. The criminal-enterprise-framing intervention (asset seizure) constrains downstream consumption — the Journal reports that proceeds fund “rented luxury cars and homes with infinity pools” and that “cash in the form of crypto is spirited back to Romania and flows to real estate, gambling and crime bosses” — but does not affect upstream builder capacity. The architectural-framing intervention (mobile wallets) would address the terminal attack surface but introduces account-takeover and SIM-swap attack surfaces that the technology-framing intervention does not. The enforcement-priority-framing intervention operates on attention rather than attack surface; its leverage is political, not technical.

The synthesis that no single component produces: Ohio’s expected 75-80% reduction in measured skimming fraud is the correct prediction for the intervention chosen, against the metric named, in the reference class identified. It is also, by the same logic, an incomplete prediction about welfare fraud as a whole. The gap between “what the chip card fixes” and “what the welfare recipient experiences” is the same gap that federal reimbursement withdrawal widened and that bot-driven account-number guessing — which the Journal reports as a rising criminal tactic alongside “running bogus transactions through cloned credit-card machines to move funds into their own accounts” — is now positioned to fill.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Pre-Mortem (Action Plan)
Imagines the plan has already failed, then works backward to find out why.
Probabilistic Forecasting
Puts calibrated probabilities on what happens next.
Wicked Problems
Treats a problem as wicked — no stopping rule, no clean test of success, every attempt consequential.