Legacy systems in government aren’t just old—they create real risk, cost, and service issues for residents.

• Security vulnerabilities: A 2019 Government Accountability Office study found that among the 10 most critical federal legacy systems, the Department of Homeland Security’s system alone had 168 critical risk vulnerabilities that hackers could exploit.
• Obsolete technology: Some core systems still run on COBOL, an outdated programming language. The Department of Education used such a system, and there are now very few technicians who can support it.
• Extreme maintenance costs: In the 2019 fiscal year, the federal government spent about $80 billion of its $90 billion IT budget just to maintain existing legacy systems, leaving relatively little for modernization.
• Service breakdowns: Old systems are cumbersome, prone to errors, and hard to adapt. For example, the Wisconsin unemployment system lacks basic capabilities like email support, document uploads, and automated changes, which slows down processing and frustrates both staff and residents.

When these systems fail or lag, people feel it directly—through delayed benefits, inaccurate records, and long wait times. That’s why replacing outdated government technology is less about “nice-to-have” upgrades and more about reimagining how essential public services are delivered.

Wisconsin’s unemployment system is a clear example of how technical debt, scale, and politics intersect to slow modernization.

• Legacy platform: The system runs on COBOL, which makes it hard to find skilled support and even harder to add modern features like online document uploads or automated updates.
• Strain under crisis: During COVID-19, the Wisconsin Department of Workforce Development processed more unemployment claims in nine months than in the entire period from 2016 to 2019 combined. The outdated system struggled under that volume, and some newly unemployed residents waited months for their claims to be resolved.
• Failed past attempts: A 2007 effort under Gov. Jim Doyle to overhaul the system was canceled when it fell behind schedule and looked likely to exceed its $24 million budget. That left the state still dependent on the old platform.
• Funding and political gridlock: Gov. Tony Evers has proposed a new modernization plan estimated at $90 million over 10 years. However, the Republican-controlled legislature and the governor have been at an impasse over funding and responsibility. A Politifact review noted that Evers cannot fully fund the project through executive action alone; the legislature must allocate money on a biennial basis.

The result is a system everyone agrees is outdated, but that remains in place because of cost concerns, project risk, and political disagreement. It shows that modernizing government tech isn’t just a technical project—it requires sustained funding, bipartisan alignment, and realistic expectations about scope and timelines.

Modernizing government technology means rethinking both the systems and the way projects are planned and funded. Based on the examples in the text, a few practical steps stand out:

• Prioritize high-risk, high-impact systems first: Focus on systems with known vulnerabilities or major service impact—like those identified by the Government Accountability Office as “most critical.” For example, systems with hundreds of critical vulnerabilities or those that are 50 years old should be at the top of the list.
• Break large projects into smaller phases: The U.S. Citizenship and Immigration Services planned to spend $500 billion to convert nearly 100 paper forms to digital, but after about 10 years had spent $1 billion and fully digitized only one form. That experience shows the risk of massive, all-at-once projects. Smaller, phased efforts with clear milestones are easier to manage and adjust.
• Align funding with long-term plans: Wisconsin’s unemployment system illustrates that governors cannot modernize alone. Legislatures need to commit to multi-year funding—such as the proposed $90 million over 10 years—and stick with it across election cycles.
• Design around user needs: Basic capabilities like email communication, document uploads, and automated changes should be standard. These features directly reduce backlogs and errors, and they make systems more resilient during spikes in demand, such as during a pandemic.
• Address political gridlock early: Disputes over responsibility and funding, as seen in Wisconsin and in broader debates over COVID-19 relief and mask mandates, can stall even well-designed projects. Building bipartisan support and clear accountability from the outset is essential.

By taking these steps, governments can gradually reshape their technology landscape—moving away from fragile, expensive legacy systems toward more adaptable platforms that better match today’s expectations for digital public services.