Keynote Speech: Payment Integrity @IBM Center for The Business of Government and National Academy of Public Administration
"Connecting the Dots: Data Integrity, Payment Accuracy, and the Public Trust"
Good morning everyone, I’m honored to be here with so many of you working at the intersection of public service, policy, and digital transformation.
Now, I’ll admit, when I told a friend I was giving a speech on “payment integrity,” they said, “That sounds... thrilling.”
I said, “It’s like true crime, but with spreadsheets.”
But the truth is, there is something thrilling about this work. Because every time we prevent an improper payment, we protect a veteran’s pension. A child’s healthcare. A family’s economic recovery. A taxpayer’s trust.
Behind every number is a name. And behind every improper payment is a missed opportunity to deliver services better, and smarter.
Let’s be real about the scope.
In fiscal year 2023, the federal government made $236 billion in improper payments. That’s not a typo. That’s nearly a quarter of a trillion dollars, and it’s not an anomaly. We've seen similar numbers year after year.
Some are overpayments. Some are underpayments. Some are flat-out fraud. Most are just... errors. Avoidable, frustrating, expensive errors.
The Social Security Administration (SSA) alone, one of our most trusted and vital institutions, manages some of the most complex benefit programs in government. And the complexity matters.
Take Disability Insurance (DI). A GAO study showed that 71% of DI beneficiaries who returned to work and earned income were overpaid. Often, for 9 months or more, and usually without realizing it.
One man in Maryland, who was recovering from a serious injury, started working part-time while still receiving benefits. He reported his wages, but due to delays in processing, and outdated systems, he was overpaid by over $11,000. He didn’t try to game the system. But when the SSA finally caught the mistake, he was on the hook. And stressed. And confused.
That story is not rare.
So, what’s going wrong?
Three big things:
• Late or missing data, especially wage data.
• Complex eligibility rules, that trip up both beneficiaries and administrators.
• Siloed systems, agencies working in parallel, but not in sync.
Now imagine this: what if wage data didn’t rely on self-reporting? What if eligibility checks weren’t based on forms mailed in or scanned weeks late? What if one agency’s knowledge could securely inform another’s decision?
That’s where data exchange becomes not just a solution, but a necessity.
Let me highlight a real success story: SSA’s Payroll Information Exchange, or PIE.
PIE connects SSA directly to payroll providers. So instead of waiting for beneficiaries to report wages, or worse, waiting until tax season, SSA gets real-time earnings data. That means fewer overpayments, faster corrections, and better service.
One SSA analyst I spoke with recently described it like this:
"Before PIE, we were trying to drive with the rearview mirror. Now we can see the road ahead."
It's still early, but PIE represents a shift from reactive to proactive, and that’s the kind of transformation we need.
Now, let’s talk about a lesser-known but incredibly powerful tool: the Numident file, SSA’s master database of identity information.
It contains every individual’s:
• Social Security Number
• Name
• Date of birth
It’s the most authoritative source of truth for identity and life status in the federal government. And yet, its use across agencies is still limited, not because it isn’t valuable, but because of the Privacy Act.
The Privacy Act of 1974 is vital. It protects personal information from misuse. It ensures government doesn’t become Big Brother.
But it also means SSA cannot share Numident data with agencies like the Small Business Administration (SBA), even when that data could prevent payments to the deceased, or flag identity fraud.
So what do we do?
We don’t go around the law. We work through it, with smart partnerships.
Enter the Department of the Treasury and the Do Not Pay (DNP) system, one of the smartest innovations in payment integrity.
DNP allows agencies, including SBA, to screen applicants against trusted data sources, including Numident, without direct access to that data.
Here’s how it works:
• SBA submits a batch of applicants to DNP.
• DNP checks them against death records, SSNs, and other federal sources.
SBA gets a result like:
• “SSN not valid”
• “DOB mismatch”
• “Deceased – verified by SSA”
No raw data. No privacy breach. Just the insights needed to stop improper payments before they happen.
It’s the perfect model for privacy-preserving data use: you don’t need the file, you just need the flag.
This model, where SSA data powers Treasury systems that protect SBA dollars, is more than just interagency cooperation. It’s a case study in what’s possible when we connect the dots.
And we should be scaling it.
We should be:
• Broadening DNP access
• Accelerating routine use agreements
• Funding better infrastructure for real-time exchange
Because these tools don’t just protect money, they protect people. They protect trust.
And that trust, between the public and its government, is everything.
At the start of this speech, I joked that payment integrity is like true crime with spreadsheets.
But let’s be honest: this work is about solving mysteries.
It’s about finding errors before they grow, detecting patterns before they spiral, and making sure every dollar we spend reflects our values, transparency, accountability, and service.
And if we do this right, if we keep connecting the dots between agencies, between datasets, and between missions, we’re not just saving money.
We’re building a smarter, more trusted government.
Comments
Post a Comment