With 3M+ people filing for unemployment benefits last week, US states are seeing their aging tech systems get overwhelmed and experts say it will only get worse
Laid-off workers struggle to file claims for unemployment benefits; economists say the situation will likely get worse. Tweets: @nxthompson and @parismartineau Tweets: Nicholas Thompson / @nxthompson : 3 million Americans filed for unemployment last week. That's more than four times higher than the worst number we've ever recorded. Unemployment offices and websites can't keep up. 5/ https://www.wired.com/... Paris Martineau / @parismartineau : many states' unemployment systems are decades old & rely on out-of-date programming langs like COBOL when 3,283,000 people filed for unemployment last week, most states' systems couldn't handle it. many are still largely inaccessible—that's a huge problem https://www.wired.com/...
Context & Ripple Effects
The claim surge Nicholas Thompson flagged — more than four times the previous record week — hit state unemployment offices running on decades-old COBOL codebases, and Paris Martineau's reporting shows the result: websites and phone lines that simply cannot process the volume. This was the first stress test of the CARES-era safety net, and it failed at the front door.
The arc since then confirms the strain wasn't temporary: gig workers reported that the new benefits in the $2T COVID-19 relief bill were hard to access and slow to arrive on the same overloaded systems, and years later the Atlanta Fed put a price tag on those delays — at least $40B in lost US GDP through end of 2020.
First-order effects
- Laid-off workers across US states could not file claims or reach benefits because COBOL-based systems built for far lower volumes buckled under 3,283,000 filings in a single week.
- State unemployment agencies faced immediate pressure to patch or staff around legacy code while economists warned filings would keep climbing.
Second-order effects
- Congress's expanded benefits for gig workers routed through the same clogged pipes, so relief money authorized in Washington sat undelivered behind broken state interfaces.
- States rushing automated fixes carried real risk: Michigan's earlier flawed system falsely charged thousands of claimants with fraud and collected millions in fines, a cautionary precedent for hasty modernization.
Third-order effects
- If the pattern holds, legacy government IT becomes a measurable macroeconomic liability rather than a back-office embarrassment — the Atlanta Fed's $40B GDP figure turns maintenance debt into a budget argument.
- Labor experts' later warning that unemployment benefits are unlikely to protect against AI job losses suggests the same under-built system will face the next shock without the upgrades this crisis exposed.
The trend: US social-safety-net delivery is shifting from neglected state-run legacy IT toward a recognized national infrastructure problem whose failure costs are now quantified in GDP terms.