The Tax System Modernization Program That Would Not Die

Retaining old technology can sometimes work out in astonishing ways.

In 1943 the destroyer USS Laffey launched from Maine and began escorting convoys across the Atlantic the following year. After supporting the D-Day landing on Utah Beach and the bombardment at Normandy, the Laffey sailed around the world to the Pacific where it was in the battle of Ormoc Bay, the landing on Iwo Jima, and the invasion of Okinawa. Along the way it was hit by torpedoes, shells, and — off Okinawa — six kamikaze planes and four bombs.

The Laffey got patched up and went on to be part of the atomic bomb tests. In the early 1950s, the ship saw action in the Korean War. In the late 1950s, it patrolled the Mediterranean during the Suez Crisis. And finally, during the Cuban missile crisis, the Laffey was on standby to evacuate naval officers in the event of an attack on Washington. In 1975 it ended up being the last Sumner-class destroyer to be decommissioned and is now a museum ship in South Carolina, where it is appropriately nicknamed “The Ship That Would Not Die.”

The Department of Government Efficiency is taking a look at the IRS’s modernization program and operations, and it has found computer systems and contracted services that aren’t quite the tax system equivalent of the Laffey.

The initial public assessment offered by Sam Corcos, the temporary DOGE senior adviser for technology and modernization at Treasury, is that the technology and contracts badly need an update. That assessment will come as no surprise to Tax Notes readers, who are keenly aware that the agency’s technology modernization efforts are well intentioned but often redolent with déjà vu.

“The IRS has some pretty legacy infrastructure. It’s actually very similar to what banks have been using — it’s old mainframes running COBOL and assembly [language code], and the challenge is how do we migrate that to a modern system?” Corcos told Laura Ingraham on Fox News on March 20. “Typically in industry, this takes a few years and maybe a few hundred million dollars. We’re now 35 years into this program,” he said, asserting that it’s also $15 billion over budget. The IRS didn’t respond to a request to verify this claim by press time.

The IRS’s computer systems should be able to interact with each other, but that’s difficult when some of them are running older languages that don’t integrate well with newer ones. Moving away from older technology gets even more challenging with time, which can result in siloed data and computational capabilities. Maintenance also becomes a challenge, as parts for older hardware — not to mention employees who know how to maintain it — grow scarce. And as history shows, turning the outdated-technology ship around is especially hard when confronted with the realities of government.

A Brief History of IRS Tech Modernization Woes

The IRS automated the tax administration system in the 1960s. According to a 1987 Government Accountability Office report, from 1960 through at least 1987 the IRS collected paper returns from taxpayers at 10 service centers around the country, compiled the information on magnetic tapes, and sent them to the National Computer Center, which maintained taxpayers’ master file accounts. (“ADP Modernization: IRS’ Redesign of Its Tax Administration System,” GAO/IMTEC-88-5FS (1987).) Accessing data internally wasn’t always as easy as it should have been, in part because of the simultaneous, but siloed, development of computer systems for functional units within the IRS that didn’t necessarily integrate with each other.

One of the more egregious examples of wheel-spinning on IT projects began in the 1980s. In 1982 the IRS launched the Tax System Redesign — an attempt to integrate automated data processing systems and make data more readily available internally — at a budgeted cost of $83.6 million. The plan was to comprehensively improve the IRS’s technology by introducing up-to-date computers and telecommunications, provide faster access to taxpayer information and better link-related information, and automate manual processes.

The GAO report explains that over the years the IRS had developed systems to support specific functions, but that those systems were never integrated “because of the design constraints of the master file processing system.” The result was that “agency users are confronted with taxpayer account data that differ in format, content, and degree of accessibility.”

The redesign effort was plagued with difficulties as the IRS searched for a feasible approach. There were leadership changes in the IRS and Treasury that resulted in changes to the plan, internal dissension about whether the redesign strategies would support IRS mission requirements, and redirections that caused delays. (“ADP Modernization: IRS’ Tax System Redesign Progress and Plans for the Future,” GAO/IMTEC-88-23BR (1988).)

By 1988, the IRS had finally approved a system redesign management plan that featured large corporate databases of information used by most of the departments, departmental databases, and systems to provide automated support for specific local offices. The goal was to implement the objectives in two phases, to be completed by 1994 and 1998.

Things didn’t go quite according to plan. In 1995 the GAO observed that the IRS had already spent $2.5 billion on tax systems modernization since 1986 but that “management and technical weaknesses must be corrected if modernization is to succeed.” The GAO added that the “IRS does not have a comprehensive business strategy to cost-effectively reduce paper submissions, and it has not yet fully developed and put in place the requisite management, software development, and technical infrastructures necessary to successfully implement an ambitious world-class modernization effort.” In 1995 the projected total for the tax systems modernization program was $8 billion.

In 1997 the National Commission on Restructuring the IRS wrote “A Vision for a New IRS,” in which it noted that the modernization program had failed “because the IRS did not have a consistent long-term strategic vision to guide the project,” resulting in a lack of technology integration. The report said the agency had drafted a new modernization blueprint to create integrated databases and break down the barriers between effectively siloed functional units and their computer systems.

The Business Systems Modernization program was initiated in fiscal 1999. According to the GAO, as of February 2004, $1.7 billion had been appropriated for the program. Then-IRS Commissioner Mark Everson told Congress the same month that several parts of the program had missed deadlines and overrun their budgets and should be re-bid. (Prior coverage: Tax Notes, Feb. 16, 2004, p. 818.) Everson said there had been modernization progress but admitted that the IRS wasn’t “at best practice yet,” and that updating the master file and computer infrastructure was still problematic.

In 2006 the Treasury Inspector General for Tax Administration assessed the results of the eight years of the Business Systems Modernization program as “mixed.” That year, the IRS switched its approach from completely replacing existing business systems to using them to accomplish the modernization process. It also decided to take over the role of systems integrator from the principal contractor.

On tax day in 2018, the IRS experienced a failure of the COBOL-based individual master file system that resulted in taxpayers being blocked from making payments and e-filing returns. (Prior coverage: Tax Notes, Apr. 23, 2018, p. 545.)

Then-acting IRS Commissioner David Kautter had the unfortunate job of testifying before the House Oversight Government Operations Subcommittee that morning. He promised to resolve the problem, and the House passed another modernization bill the same day.

Another modernization program, the Integrated Modernization Business Plan — a “six-year road map for achieving necessary modernization of IRS systems,” according to the chief information officer — was released in 2019. (IRS, “IRS Integrated Modernization Business Plan” (Apr. 2019).) The chief information officer’s transmittal message repeated the common theme of IRS efforts: “Our legacy computing infrastructure cannot keep pace with the desire for instantaneous data, real-time interactions, and other customer-centric services,” adding that “the cost to operate our current technology ecosystem continues to increase.”

The Inflation Reduction Act ultimately gave the IRS approximately $57.8 billion, after Congress trimmed the $80 billion it initially allotted, to update antiquated computer systems, improve taxpayer services, and increase compliance and enforcement of high-income taxpayers and large corporations.

In an October 2024 update on modernization efforts following the IRA, TIGTA noted that “the IRS continues to maintain some of the oldest information technology systems in the federal Government.” (TIGTA, “Major Management Challenges Facing the IRS in 2025” (Oct. 15, 2024).) The same report said that $1.6 billion of the $4.8 billion total for business systems modernization in the IRA had been spent by June 2024, and seven of the 13 milestones for “delivering cutting-edge technology, data, and analytics to operate more effectively” had been completed. TIGTA also pointed out that the IRS has historically run into problems implementing effective security and audit logs when introducing new technologies.

On the bright side, TIGTA reported that the IRS was working to streamline its data processes and efficiency, including by publishing an Enterprise Data Catalog powered by artificial intelligence that “provides a comprehensive view of data assets and should enable data sets to be standardized so they can be used for efficient discovery, understanding, and lineage tracking.” The IRS also added data assets to the Enterprise Data Platform, allowing universal data access and analytics for users.

The Challenge for DOGE

As in any organization, technology at the IRS requires regular updating, so there is some excuse for the seemingly perpetual work of IRS modernization. But actually updating the technology seems to be less of a challenge than the development and implementation of the IRS’s projects.

Treasury Secretary Scott Bessent blamed “entrenched interests” for siphoning off tax dollars through contracts. Corcos said he found the disconnect between the IRS’s leadership and the IT employees surprising. “It doesn’t take a lot, just somebody who cares, to solve these problems,” he said. The IRS also has to contend with Congress, which has yanked funding for projects in the past.

Inertia is another problem, according to Corcos. He said that DOGE had found contracts that it couldn’t identify the purpose of and canceled them. DOGE’s website lists multiple IT-related contracts that have been terminated or partially canceled. Corcos also said that DOGE has cut projects from the IRS’s modernization budget that would have increased complexity in the agency’s systems. Tax Notes was unable to find publicly available contact information for DOGE to request further information about the terminated contracts.

Corcos compared the IRS’s IT department to that of a typical midsize bank, which he said would have an IT staff of 100 to 200 and an operation and maintenance budget of roughly $20 million, while the IRS’s IT department has 8,000 employees and a budget of $3.5 billion. The IRS’s most recent Data Book shows roughly $3.5 billion in costs incurred for information services in fiscal 2023. (IRS Data Book 2023, Pub. 55-B, at 73, Table 32 (2024).) “I don’t really know why yet, but I will tell you that 80 percent of that budget goes to contractors and licenses,” he added. Corcos said that the IRS has “a lot of budget” for IT.

Modernization projects at the IRS might be challenging, but they ought to stop bearing such a striking resemblance to the interminable Dickensian case of Jarndyce v. Jarndyce. Projects need not drag on year after year, with relatively little to show in the way of a conclusion. “I’m actually pretty optimistic that we can solve this,” Corcos said, adding that his experience had been “very positive” working with the IRS’s IT employees to try to solve the problems that they know need to be solved, but “they’ve been in this situation where their hands are tied.” He added, “One encouraging thing is we actually have quite a lot of software talent on the ground [in the IRS], the people writing code.”

Bessent said his priorities for the IRS are collections, privacy, and customer service, and that “none of those are being well served” by the current state of the modernization project. Protecting security and maintaining privacy has been a stumbling block for the IRS’s past attempts to address aging infrastructure. Bessent also promised to bring more DOGE advisers out to do interviews so that taxpayers could “see what’s being done.” Echoing decades of reports on IRS modernization projects, Bessent expressed the hope that the agency would soon work “better, cheaper, faster, and with more privacy.” It’s a fine vision, but the real test will be whether the IRS can bring it to fruition.