The $2.8 Trillion Question Nobody's Answering: When Grandma Can't Pass Fraud Checks and Criminals Walk Free

GM, Welcome Back to the Dead Drop.

Let me ask you a question: if I told you the federal government has made $2.8 trillion in improper payments since 2003, what's the first word that comes to mind?

If you said "fraud," you're wrong.

But you're not alone. Politicians, news anchors, and angry taxpayers all make the same mistake, conflating two very different problems into one convenient villain. And that confusion is exactly why we're hemorrhaging billions annually while legitimate beneficiaries get locked out of the programs they've earned.

I spent two decades hunting criminals across four continents for the U.S. Government. I know what fraud looks like: it's the organized crime ring that created fake clinics and bilked Medicare for $100 million. It's Feeding Our Future's $250 million pandemic scheme using shell companies and fabricated meal counts. It's willful misrepresentation for financial gain.

But here's what continues to bother me: the homeless veteran who can't access his disability benefits because he doesn't own a smartphone for identity verification. The 85-year-old stroke victim whose family can't prove his identity through a video call because he's blind and hooked up to machines in a hospital bed. The grandmother whose Social Security payment gets flagged as suspicious because she doesn't have the technical know-how to pass a biometric liveness detection test.

These aren't fraud. These are what the Government Accountability Office calls "improper payments," and the distinction matters more than you think.

The Definitions That Destroy Context

Improper payments are any payments that shouldn't have been made or were made in the wrong amount. Period. That includes overpayments, underpayments, payments made without proper documentation, and payments to ineligible recipients.

Fraud is obtaining something of value through willful misrepresentation. It requires intent, deception, knowledge of wrongdoing.

Here's the critical distinction the GAO hammers home repeatedly: while all fraudulent payments are improper, not all improper payments are fraud. In fact, according to recent testimony, only a small portion of the $162 billion in FY2024 improper payments involved actual fraud; but this is still an ENORMOUS problem.

Let me show you what this looks like in reality:

Improper Payment: A Medicaid provider bills for 40 hours of childcare but the attendance records only document 30 hours. The payment goes through for the full 40. That's a $231 million problem in Minnesota's childcare system according to the May 2025 HHS Inspector General audit, with 11% of payments containing attendance-related errors.

Fraud: Feeding Our Future submits completely fabricated meal count sheets and invoices for children who were never fed, using shell companies to launder $250 million in administrative fees. That's 70+ defendants facing federal charges.

See the difference? One is administrative failure, documentation gaps, honest mistakes in a complex system. The other is organized criminal enterprise.

But here's where it gets ugly: 77% of Medicaid's improper payments stem from insufficient documentation. Not criminal intent. Missing paperwork.

The Political Pressure Cooker: Why Agencies Choose "Pay and Chase"

Now let me tell you about the incentive structure that created this disaster.

Imagine you're running a benefits agency. You deny someone's claim because you can't verify their identity or their paperwork looks suspicious. That person calls their congressman. The congressman, looking to deliver constituent service and carry political favor, sends your agency a very nasty message demanding to know why you're denying benefits to deserving Americans.

This happens hundreds of times. Thousands. Your agency leadership starts to notice a pattern: when you approve questionable claims, nobody complains. When you deny suspicious ones, you face congressional inquiries, media scrutiny, angry phone calls.

The rational response? Optimize for approval rates, not fraud detection.

This is how you get what the industry calls "pay and chase": pay out the benefits first, chase down the fraudsters later. It's the exact opposite of prevention.

The Obama Administration recognized this problem back in 2010 when HHS, CMS, and DOJ announced they were "shifting from a 'pay and chase' approach toward fraud prevention." That was 15 years ago. We're still chasing.

Why? Because prevention creates visible victims with congressional representatives. An elderly constituent denied benefits makes headlines and phone calls. A fraudster who never got paid in the first place stays invisible (sometimes).

The Congressional hearings on improper payments are crystal clear on this dynamic. As one House Oversight testimony put it: "The best way to reduce improper or fraudulent payments is to not make them. Preventive controls, as their name implies, are meant to stop improper and fraudulent payments before they occur."

Sounds obvious, right? But prevention requires saying "no" to people, and agencies are politically incentivized to say "yes."

The Grandmother Problem: When Fraud Checks Become Access Barriers

Now here's where the system breaks catastrophically.

Modern fraud prevention relies heavily on biometric verification: facial recognition with liveness detection, 1:1 matching (is this photo the same person as your ID), 1:n matching (is this person in our database of known fraudsters).

The technology is sophisticated. Passive liveness detection analyzes skin texture, light reflections, micro-movements to distinguish a real person from a photo, video, deepfake, or 3D mask. It can spot synthetic identities. It stops presentation attacks.

It's also completely inaccessible to a huge swath of legitimate beneficiaries.

ID.me, the identity verification service used by the VA and other agencies, generated over 700 complaints in just four months (October 2021 to January 2022). The stories are harrowing:

An 85-year-old veteran having a stroke, blind and unable to speak, lying in a hospital bed. His child tries to access his medical records. ID.me requires a video verification call. "He's laying in a hospital bed in a hospital gown, blind, unable to talk, hooked up and drugged up," the family member wrote to the VA helpline. "This is far too personally intrusive and undignified."

After uploading documents, answering questions, submitting a selfie, ID.me wouldn't validate the veteran's identity. Why? Because "the law prevents him from validating identity with anyone but the individual," even though the mother had power of attorney.

Or consider homeless veterans, 51,936 of whom the VA permanently housed in FY2025. These are people who often lack smartphones, stable addresses, or the technical literacy to navigate biometric verification systems. A GAO report found they face barriers in obtaining even basic state IDs: no money for fees, no supporting documents, no residential address to provide.

And here's the cruel irony: these are the exact populations most likely to have their benefits flagged as suspicious. Homeless veteran using a shelter address? Flagged. Elderly person without a smartphone? Can't complete verification. Someone with limited English proficiency trying to navigate a video call? Denied.

The same fraud prevention tools designed to stop criminals are locking out the most vulnerable legitimate beneficiaries.

The DOGE Effect: When You Fire the Fraud Fighters

Now let's add one more disaster to the pile: what happens when you gut the workforce responsible for verification?

Between January and November 2025, the federal workforce dropped from 3.015 million to 2.744 million workers. That's 270,000 people gone. DOGE's "efficiency" drive resulted in massive cuts across agencies responsible for benefits administration.

The Social Security Administration, which administers programs serving millions of Americans, saw 2,477 employees accept voluntary separation by mid-March 2025. The agency was already understaffed before DOGE announced plans to close 47 field offices.

Here's what that means in practice: longer wait times for in-person verification. Slower processing of appeals when automated systems deny legitimate claims. Reduced capacity to conduct the manual reviews that distinguish the struggling grandmother from the organized fraudster.

Treasury's "Do Not Pay" system, which prevents payments to deceased individuals and ineligible recipients, desperately needs permanent access to Social Security's full death data. The pilot program prevented $109 million in improper payments in just one year. But the authorization expires in December 2026 unless Congress acts.

Meanwhile, DOGE laid off 350 people from the National Nuclear Safety Administration who oversee nuclear arsenal safety. When you're firing critical personnel with no understanding of what they do, you're not improving efficiency. You're creating cascading failures.

The Impossible Choice: Fraud Prevention vs. Access

Here's the operational reality agencies face right now:

Deploy sophisticated biometric fraud controls that lock out elderly, homeless, and vulnerable populations who can't pass technical verification. Accept the congressional complaints and media stories about denied benefits.

OR

Loosen verification requirements to ensure access, revert to "pay and chase," and accept that organized crime will systematically exploit the system for billions.

There is no good choice under current constraints. You either deny benefits to people who've earned them, or you hemorrhage taxpayer money to criminals.

The Minnesota childcare situation illustrates this perfectly. An 11% error rate in payments, totaling $231 million, with most errors involving attendance documentation. Real fraud exists (Feeding Our Future proved that). But the viral video claiming "empty daycares everywhere" led to payment freezes affecting 19,000 children across Minnesota, most receiving legitimate services.

State inspectors confirmed 8 of 9 centers featured in the video were operating normally. But the political theater got headlines. The actual fix, automated attendance tracking with real-time electronic reporting, got ignored.

The Fraudfather Bottom Line

We're asking the wrong question.

The debate isn't "why are there improper payments?" The answer is obvious: because perfect fraud prevention is impossible without creating equally impossible barriers for legitimate beneficiaries.

The real question is: "How do we build verification systems that can distinguish grandma from the fraudster without requiring grandma to own a smartphone, complete a video call, and pass biometric liveness detection?"

Here's what would actually work:

Make SSA's death data sharing with "Do Not Pay" permanent. This single change prevented $109 million in payments to deceased individuals in one year. Congress keeps letting the authorization expire. Stop it.

Fund automated attendance and eligibility verification systems. Real-time electronic reporting catches documentation errors before they become improper payments. Minnesota's childcare system desperately needs this. So does every other benefits program.

Create alternative verification pathways for vulnerable populations. Not everyone has a smartphone. Not everyone can complete a video call. Build systems that work for homeless veterans, elderly beneficiaries, and people with disabilities without sacrificing fraud detection.

Staff agencies to handle manual review capacity. Automated systems will always flag false positives. You need human investigators to tell the difference between suspicious activity and struggling beneficiaries. DOGE firing 270,000 federal workers destroyed this capacity.

Stop optimizing for approval rates. Political pressure drives agencies to say "yes" by default. That creates the pay and chase model. If we want prevention, Congress needs to stop punishing agencies for denying suspicious claims.

The $2.8 trillion in improper payments since 2003 isn't a fraud problem that enforcement theater can fix. It's a systems design problem that requires actual investment in verification infrastructure.

Right now, we're doing the worst of both worlds: denying benefits to legitimate recipients who can't pass biometric checks while organized crime walks through documentation loopholes we refuse to close.

That's not fraud prevention. That's fraud facilitation with a side of cruelty to the vulnerable.

When you build systems that can't tell grandma from the fraudster, everyone loses except the criminals.

6,000+ fraud investigators, executives, investors, and skeptics read this newsletter because they refuse to be victims. While most people argue about headlines, we track the institutional fraud that creates rationalization frameworks for crime at every level. The Dead Drop connects monetary policy to your portfolio, political theater to who pays the bill, and systemic theft to why everyone's becoming a scammer. We don't cover symptoms - we expose the fraud infrastructure that makes honest people feel like suckers. Know an investigator, executive, or operator who needs this intelligence? Forward this.

$68M Brooklyn Fraud Shows Why "Middleman" Models Are Criminal Playgrounds

Two Brooklyn recruiters, Manal Wasef and Elaine Antao, just pleaded guilty to exactly the kind of fraud I'm talking about in this week's piece: organized criminal enterprise masquerading as legitimate healthcare.

Between 2017 and 2024, they billed Medicaid $68 million for adult day care and home health services they never provided. Their scheme was elegant in its simplicity: pay kickbacks directly to Medicaid recipients, use their enrollment to generate fraudulent claims, launder the proceeds through shell companies, rinse and repeat for seven years.

This is the sixth and seventh guilty plea in an eight-defendant case. That's not improper payments from documentation errors. That's systematic theft with multiple co-conspirators.

Here's what keeps this fraud alive: the Consumer Directed Personal Assistance Program (CDPAP) allows people with minimal healthcare experience to care for elderly or disabled relatives, with hundreds of "middleman" firms acting as payroll agents between caregivers and Medicaid. Minimal oversight.

Notice the pattern? The fraud isn't happening because criminals are sophisticated. It's happening because we built a system with intentional gaps. Middleman firms with minimal oversight, payment based on enrollment rather than verified service delivery, no real-time attendance tracking.

The same $231 million Minnesota childcare problem. The same structural vulnerability. We pay first, verify never, chase eventually.

Wasef and Antao face maximum ten years. They stole $68 million over seven years. That's $9.7 million annually. Even if they serve the full decade, the return on investment for organized crime is staggering.

This is why "pay and chase" doesn't work. By the time you catch them, they've already won.

Detroit Student Aid Fraud Shows How Pattern Detection Beats Manual Review

Michelle Denise Hill ran a textbook fraud operation for a decade before anyone caught her. From 2015 to 2025, she enrolled roughly 80 people in Wayne County Community College using stolen identities, submitted fraudulent FAFSA applications, and routed $2.5 million in federal student aid refunds back to herself through at least seven bank accounts.

Here's what makes this case instructive: Hill didn't just steal the money and run. She maintained operational sophistication. She completed the coursework herself (agents found 800+ course files), controlled victims' debit cards because they "trust her more than they trust themselves," and even handled communications with professors while impersonating students.

The scheme only unraveled in 2022 when a school official noticed a cluster of aid applications all listing the same Florida high school. That single observation triggered an investigation that revealed shared challenge-question answers, identical addresses, repeated phone numbers, and Hill's name appearing as a personal reference across multiple supposedly unrelated applications.

Ten years. Eighty victims. $3.2 million awarded. All caught by one alert administrator noticing a pattern.

This is exactly the kind of fraud that automated analytics should catch in weeks, not decades. Banks use these same techniques to detect money laundering networks. Credit card companies spot fraud rings instantly. The technology exists to flag when multiple "students" share IP addresses, banking details, and reference contacts.

But federal student aid systems still rely on manual review and post-payment investigation. By the time Hill's pattern became obvious enough for human detection, she'd already successfully routed millions through her operation.

This wasn't an improper payment from documentation errors. This was systematic criminal enterprise that exploited our refusal to deploy basic fraud detection infrastructure.

The Machiavelli Minute: Rule #1 - Systems Optimize For Survival, Not Success

Every organization tells you what it's supposed to do. Performance metrics. Mission statements. Strategic goals. Ignore all of it. Watch what behavior gets punished versus rewarded, and you'll understand what the system actually optimizes for.

Benefits agencies don't optimize for fraud prevention. They optimize for avoiding congressional complaints. When you approve a questionable claim, nobody calls. When you deny a suspicious one, you get angry constituents, media scrutiny, and legislators demanding answers. The rational move? Approve everything, chase fraud later.

This isn't incompetence. It's survival.

Minnesota's childcare system doesn't optimize for stopping fraud. It optimizes for maximizing enrollment and access. Legacy rules required upfront payment based on enrollment, not verified attendance. Why? Because verification creates barriers, barriers reduce enrollment, reduced enrollment generates political backlash. The system did exactly what it was designed to do: prioritize access over accuracy.

The fraudsters understood this perfectly. They didn't break the system; they exploited what it actually optimized for.

DOGE learned this lesson the hard way. Fire 270,000 federal workers to "cut waste," but suddenly agencies can't process benefits, handle appeals, or distinguish legitimate claims from fraud. Turns out those "inefficient bureaucrats" were the manual review capacity preventing system collapse.

Here's the rule: institutions will always sacrifice their stated mission to preserve their political survival. The agency that fights fraud aggressively dies in the next budget cycle. The agency that approves everything and apologizes later survives.

Want to predict system behavior? Don't read the mission statement. Find out what triggers congressional inquiries, and you'll know exactly what gets approved.