Obsolescence Over Modernization Part I: The Growing Crisis in Government Systems
The Trump administration has made clear its intentions to rein in government and eliminate waste, as noted in the “Cost Efficiency Initiative” executive order (EO) calling for “a transformation in Federal spending on contracts and grants.” While the EO is sweeping, updating legacy systems that underlie the operations of many federal agencies should be a priority. From tax collection and benefits disbursement to managing federal assets and the government workforce, federal agencies struggle under the burden of outdated technologies.
Managing federal government operations is a substantial undertaking. Annual federal information technology (IT) spending exceeds $100 billion. Still, up to 80 percent of the budget is dedicated to the operation and maintenance of existing systems, which often run on decades-old infrastructure and outdated software. These aging systems generate indirect costs as well, as agencies resort to workarounds and manual procedures to maintain their viability. Federal IT acquisitions and management have been on the Government Accountability Office (GAO) High-Risk List since 2015, as maintaining aging systems leaves only minimal funding available for modernization efforts.
Despite the GAO’s warnings of the risks posed by federal IT systems over the past decade, the issue has yet to be adequately addressed. As the Trump administration continues to champion a leaner, more efficient government, it is important to explore the root causes of this persistent issue and solve the federal IT system architecture crisis.
This first installment in our two-part series on IT modernization across the federal government identifies and evaluates current challenges, particularly the institutional barriers and economic incentives that hinder effective reforms. The next installment will propose solutions and reform strategies to overcome these obstacles and build more cost-effective, secure, and efficient IT systems.
The Challenge of Legacy Systems: A Crisis of Obsolescence
Legacy systems—some predating the modern internet—continue to perform basic federal government functions, such as tax collection, benefits management, and human resource management. Through decades of incremental upgrades, collections of legacy hardware and software systems evolved into complex and unique patchworks, customized and integrated with a variety of proprietary technologies. The non-standard nature of these enterprise IT systems makes it exceedingly difficult to modernize or integrate them into newer systems, as their legacy roots are so vastly different from modern architectural designs.
As a result, federal agencies have lagged behind private-sector firms in modernizing their operating infrastructure. Public-sector operations and maintenance budgets can be significantly higher than their private-sector counterparts, whose operations and maintenance costs are typically under 50 percent. The failure to innovate generates unique challenges. For instance, one legacy aspect of many government systems is the use of older programming languages like Common Business-Oriented Language (COBOL), introduced in 1959 by Grace Murray Hopper. While older programming languages do have benefits and are still in use in the private sector (particularly in financial institutions), it is difficult to find IT specialists familiar with legacy software.
When employees are required to learn outdated skills not typically viewed as beneficial to their careers, retention becomes a challenge. Additionally, because such patch-heavy Franken-systems often rely on custom workarounds to overcome limitations, their maintenance costs are substantial. They also have high error rates and present significant cybersecurity risks—unacceptable outcomes for IT systems that store and process sensitive government information. Furthermore, they are prone to data siloing, which presents data fragmentation and interoperability challenges that hinder comprehensive analytics and effective decision-making. Data silos also limit the possibility for federal agencies to explore the promise of data-intensive technologies like artificial intelligence (AI).
Institutions, Information, and Incentives
Although newer technologies and AI-assisted programs could substantially lower costs and reduce cyberthreats, many agencies fail to make the transition to more modern and cost-effective systems. In fact, adoption rates within the public sector are considerably lower than that of their commercial counterparts. Much of this gap is attributable to institutional differences between the government and the private sector rather than a lack of will or expertise to adopt newer, more efficient technologies.
Competitive markets drive superior outcomes through price signals that efficiently allocate resources to their highest valued use. Prices reflect critical information about demand, scarcity, and the costs of production that guide the actions of consumers and producers. Importantly, in the private sector, a residual claimant earns a profit or loss based on the firm’s performance; in turn, employees are rewarded based on their contribution to the company’s revenues. This creates incentives that drive efficiency, pushing firms to adopt new technologies that can lower costs and increase output.
On the other hand, federal agencies have no residual claimant and no price signals to convey the information necessary for efficient resource allocation. Rather than profits, broad public-policy objectives, political processes, and extensive stakeholder negotiations inform public-sector decisions, all of which occur in annual budget cycles and through changes in political leadership. Consequently, the competition between innovation and political and fiscal objectives (as well as other agency priorities) make it more difficult for the public sector to modernize.
A substantial body of literature examines decision-making in the public sector. One George Mason professor’s pioneering work on bureaucracy suggests that public-sector employees, like those in private firms, are rational, self-interested actors who maximize utility. Absent profits, this means maximizing the department’s budget size, number of employees, and power. With respect to legacy systems, this may lead to resistance to modernization if new technologies result in reduced departmental size or lower budgets. In the seminal work Bureaucracy and Representative Government, another scholar expounds on the notion that bureaucrats are budget maximizers and that a department head’s optimal strategy is to secure the largest budget possible for their department. In terms of IT systems, this may create incentives to maintain legacy systems with large operation and maintenance budgets—even if upgrading the system would be more efficient. At the same time, the IT department functions as a sole supplier for the agency, with information asymmetries providing an advantage in negotiations related to costs, budgeting, and the pace of modernization.
Finally, an economist provides a categorization of bureaucratic behavior in Inside Bureaucracy. Two important categories are “conservers” and “climbers.” Conservers are risk-averse and seek to maintain their income and security; consequently, they are strong advocates for the status quo. Climbers strive to maximize their utility through promotions and programs that enhance their relative status. Conservers would view changes to legacy systems with skepticism, delaying the adoption of new technologies. Climbers would support new systems based on the scale and visibility of the project. This may result in expensive new projects rather than those that most effectively address the need for modernization.
Federal agencies are institutionally distinct from firms in a market setting. Actors continue to respond rationally to incentives, but those incentives are different and yield inferior outcomes with respect to resource allocation. At the same time, information does not flow efficiently, yielding outcomes based on less-than-perfect knowledge that often favor one party over the other.
The Fallacy of Sunk Costs
In addition to different incentive structures in the public sector, the sunk cost fallacy helps explain the persistence of outdated legacy systems in federal agencies. From an economic perspective, sunk costs refer to outlays that have already been made and cannot be recovered. As such, they should be irrelevant to decisions about future outlays, which should be evaluated only in terms of future costs and benefits. Yet, behaviorally, most decision-makers find it difficult to ignore the past, giving previous investments in outdated systems irrational sway over future decisions.
Billions of taxpayer dollars have been invested in legacy systems dating back decades, with most funding going to updates and patches to keep them running. Annual appropriations have kept these systems alive, and both agencies and their congressional oversight committees often view efforts to discontinue or replace existing systems as a “waste” of the substantial previous investments. Every year that additional resources are allocated to legacy systems only makes the waste appear greater. Change has become increasingly difficult, even in cases where continued operation and maintenance costs exceed the cost of a more robust and modern system. A particularly illustrative example of this sunk-cost thinking can be seen in Internal Revenue Service (IRS) efforts to modernize its Individual Master File (IMF) system, which dates back to the 1960s and was highlighted as a key risk by the GAO in 2018. The replacement system, Tax Account Management-Individual, upgrades the software from a 66-year-old programming language to a 30-year-old programming language and runs parallel to the legacy system. The IRS expects to retire the legacy IMF in 2028, marking the end of a 19-year modernization effort.
The human capital at risk exacerbates the problem, with a specialized workforce devoted to the operation and maintenance of existing systems as strong advocates for the status quo. Agency investments in training employees to support outdated systems are also a sunk cost that distorts decision-making about investments and training on more effective modern systems.
The Entrenched Hand: Vendor Lock-In
While institutional constraints may limit the demand for updating legacy software, supply-side effects also pose a challenge to modernizing outdated systems. Vendors and providers of the legacy software systems that dominate federal agencies are firmly vested in the status quo, generating substantial revenue through maintenance and support contracts. A business model based on incremental updates and custom patches creates incentives to maintain outdated systems rather than compete for new contracts with updated systems and software that may render their services obsolete.
These incentives can result in “vendor lock-in,” making it difficult for even reform-minded agencies to update their systems. Ultimately, this exacerbates the sunk cost fallacy as agencies become so dependent on existing vendors and their products that switching costs appear prohibitive. Lock-in can hamper competition and generate significant costs. One study found that Microsoft and Oracle win at least 25 to 30 percent of their respective federal contracts non-competitively. In one instance, an agency paid an additional $112 million to purchase Microsoft Office rather than switch to Google Workspace.
Federal vendors and suppliers actively lobby and cultivate relationships within agencies to secure favorable procurement policies and practices. The specter of fear, uncertainty, and doubt is often used to restrict the adoption of modern alternatives—particularly cloud and open-source solutions—while reinforcing superficial modernization patches that continue to extend legacy systems rather than truly modernizing them. This status-quo strategy ensures continued revenue streams for existing vendors while keeping agencies dependent on outdated, inefficient technology that struggles to keep pace with evolving needs. It also creates a powerful institutional inertia that locks federal agencies into costly and inefficient legacy IT infrastructure.
In addition to hardware and software vendors and suppliers, the IT landscape is home to a plethora of service contracts. Not only does this compound the effects of vendor incentives (as private-sector employees are held in scope by the terms of their contracts), it also stands in the way of public-sector employees’ ability to innovate and experiment with in-house data. While they may be more able to recruit and retain top talent, IT service providers tend to become conservers by default, especially when paired with weak public-sector oversight.
Exaggerated risks and inflated costs of migration often dampen incentives for change, with past government IT failures used as cautionary tales to maintain the status quo. Perhaps the most notorious example of modernization gone wrong is the Phoenix pay system, adopted by the Canadian government in 2011 to streamline its 40-year-old legacy federal payroll system. Due to design flaws, a lack of testing, insufficient training, and poor management, the system was a troubling failure, resulting in overpayments, underpayments, or non-payments that adversely impacted 80 percent of the Canadian federal workforce and cost taxpayers more than $5.1 billion to resolve. Summing up the Phoenix pay system, Canada’s auditor general called it an “incomprehensible failure.”
Over the Horizon
Unlike the private sector, which is driven by profit and market share, the public sector seeks to maximize social welfare, shaped by political and budgetary cycles. Annual appropriations limits planning to 12-month cycles, fostering a “use it or lose it” mentality that can lead to inefficient spending. Continuing resolutions exacerbate the problem by freezing spending, prohibiting new program initiation, and delaying new system procurement. Furthermore, political cycles introduce policy volatility and leadership transitions that can disrupt long-term projects and reinforce risk aversion within bureaucratic structures, as seen in the persistent IT modernization failures at agencies like the Department of Defense, the Department of Housing and Urban Development, and the IRS, which continue to struggle with outdated systems due to inconsistent funding. The effects of the annual funding cycle and the disruption of continuing resolutions were seen in the U.S. Army’s attempt to adopt cloud infrastructure in 2019, which was put in jeopardy by Congress’s inability to pass a budget. This prevented the Army from funding new projects and delayed its broader IT modernization efforts until a budget was passed.
Road to Reform
Modernizing federal IT systems requires strategic planning coupled with sufficient and sustained funding mechanisms. A “whole of government” IT strategy built around mission and customer outcomes can reshape IT systems to deliver needed services more effectively and efficiently. This requires agency-level planning to evaluate broad, cross-functional initiatives that serve the agency more effectively. Predictable and sustainable funding for longer-term IT investments is also essential for multi-year modernization programs. Congress created the Technology Modernization Fund to serve this purpose. While not all projects have realized projected savings, the program has allowed some agencies to implement changes that have reduced costs. Other budgetary reforms may also be helpful, such as expanding the use of multi-year or no-year appropriations and targeted biennial budgeting to overcome short-termism.
Just as critical as implementing the right systems and software is the need for institutional changes to avoid vendor lock-in and risk aversion. Simply replacing old hardware and software will not succeed if institutional constraints and incentives continue to favor the status quo. It will be essential to conduct comprehensive inventories to assess current capabilities and identify gaps while establishing clear performance metrics and accountability measures for adopting modern technologies, such as cloud computing and agentic AI models.
Conclusion
More than a mere technical challenge, successful modernization requires institutional transformation. Ultimately, federal IT modernization requires both leadership and commitment from the executive branch as well as congressional support to ensure the resources necessary to achieve agency targets for modernization. Current efforts to right-size government offer an unprecedented opportunity to overcome institutional inertia, perverse incentives, and entrenched interests, thereby breaking the cycle of legacy funding at the expense of investing in modern infrastructure. These opportunities will be examined in more detail in Part 2 of this series.