Federal Express was able to work out a settlement with the Internal Revenue Service late last month in a long-running case involving tax credits for research and development, and the case could have broad implications for other companies claiming R&D tax credits for internally developed software.

In the late 1990s, FedEx invested heavily in overhauling its software for keeping track of the packages its customers were shipping and claimed research tax credits on much of the software development expenses.

“If somebody claims a research credit for internal-use software, though, there are additional requirements,” said David Culp, a director in KPMG’s Washington National Tax Practice. “Congress added those in 1986. It was a bit unspecific about what those requirements would be, but there had been some general agreement. It needed to be a bit more innovative, involve some more economic risk, and not be something off the shelf. FedEx thought they had satisfied this.”

He pointed out that the IRS has been working since 1986 to develop standards and regulations for claiming R&D credits on internal-use software, but they were never finalized. Not that the IRS didn’t try.

“They made two efforts,” said Culp. “One was in very early 2001. They issued some final regulations, but most of that, as it dealt with internal-use software, went along with what had traditionally been thought to be the test, the so-called ‘high threshold of innovation test.’ But they also made some additional standards dealing with the rest of the research credit. They wanted these regulations to say that any qualified research, whether it’s for software or anything else, had to be ‘undertaken to obtain knowledge that exceeds, expands or refines the common knowledge of skilled professionals in a particular field of science or engineering.’ To many practitioners, this just seemed to be a requirement that whatever you did had to be something totally new to the world, not just to a particular company. That was not how most practitioners thought the research credit was intended.”

That regulation was criticized, so the IRS tried to come up with another definition. “They took another try and proposed regulations,” said Culp. “They did away with the ‘exceeding knowledge of skilled professionals’ test, what we often call the discovery test. But they decided for internal-use software they would tighten up the additional test that had been required. Instead of looking at innovation in terms of something that makes software faster or more productive, they said that any internal use software would qualify, but it would need to be intended to be unique or novel, to be a significant and inventive [departure] from prior software. Many practitioners thought the IRS was taking out the general discovery test they had earlier for all qualified research and imposing it specifically just for internal-use software. This was heavily criticized as well.”

A few years later, the IRS did issue final regulations dealing with qualified research overall, Culp noted, but left out all the provisions dealing with internal use software. “They didn’t specify a test. At the same time, they said, ‘OK, we are going to continue working on this issue. We know it’s important to taxpayers. It is vital. It is a big part of our economy. Taxpayers do deserve some guidance. We will be working on additional regulations.’ Until then, the IRS said, taxpayers may rely on either of those two earlier formulations of what the requirements would be—this traditional high threshold of innovation test, the discovery test, and everything else that was in those regulations, or this modified standard of what this innovation test should be with this unique or novel requirement.”

FedEx filed research credit claims for the years it had been developing the package-tracking software, but decided the new tests did not apply. Culp explained it this way. ‘They said, ‘We don’t have to meet any of these tests. We don’t believe as a matter of law that the IRS has any right to impose these standards. We should still be looking at what has traditionally been considered to be the test.’ And there is support for that. Congress had expressed intent back in 1986 when they changed the law that the traditional test, including the traditional test of innovation, should be the test for internal-use software.”

The IRS denied FedEx’s claims, and FedEx filed suit against the IRS in U.S. District Court. In 2009 the court issued a memorandum and agreed with FedEx that the IRS did not have the authority to impose a higher standard, according to Culp. “If the IRS wanted a higher standard, they would have to issue specific regulations dealing with that, and they could not just say, ‘We’ll apply these other tests because we said for a while that this is the test that should apply.’ The IRS just doesn’t have that authority to impose this. It took a while after that, but the District Court entered its final decision in this case back in February of this year. They found in favor of FedEx, and they gave FedEx most of the research credit it had claimed and the IRS had denied. But they recognized that the government could appeal this case.”

The government did just that and filed an appeal in the Court of Appeals for the Sixth Circuit. “Then the parties went into settlement negotiations, so from February until basically a week ago they’ve been working out a settlement, and on August 29 they agreed to settle, that the government drops its appeal and FedEx gets whatever it claims,” Culp said last week. “We don’t know all the details by any means. It’s just pure speculation about what else they might have agreed to, but definitely FedEx was entitled to its research credit claim, and the IRS can’t pursue this research credit issue against FedEx anymore.”

Culp does not believe the case will set a legal precedent for similar cases, but he does think it will give companies more incentive to file research credit claims for internal-use software.

“For most purposes this is not any precedent,” he said. “It is binding for FedEx, but it’s just a district court opinion. Other district courts and courts of appeals are not by any means required to agree with this opinion, so the IRS is capable of going after other taxpayers disagreeing with their standards and trying to impose their higher standards in other litigations and IRS exams.”

However, he pointed out that even though it’s just one district court, it’s the only court in recent years that has ruled on internal-use software since the IRS has attempted to promulgate regulations. “We think the ruling is pretty well reasoned,” he said. “We think other courts would think about this if they had to consider the issue. A well-informed district court judge wrote what we think is a pretty persuasive decision that the IRS can’t impose other standards. We think that if the other taxpayers had this issue in an IRS exam, they could go to IRS appeals and probably get a fairly good settlement. We think IRS appeals would take this one specific decision into account very heavily in deciding how to deal with other taxpayers.”

Christine Kachinsky, credits leader for KPMG’s East Region, emphasized that while this is the first case since the IRS issued its final regulations in 2003, there are earlier cases dealing with internal-use software, just not recently. “This is the first and only one in the more recent years that deals with more recent standards that apply,” she said.

She believes companies that are investing significant dollars in their internal IT infrastructure that previously may not have claimed research credit on those because the internal use software standards were too high may now want to revisit those expenditures in light of the FedEx development. “There could be research credit opportunities there if it is approached with a level of diligence and well documented,” she advised.

The standard four-part test that is applicable to all research credit claims would need to be documented, including the technological challenges that the company faced with respect to the software development effort, what alternatives were evaluated in order to resolve those technological challenges, and what risks were associated with those challenges. “Then in relying on this lesser standard now applicable from FedEx, I think it’s important for companies to consider quantifying the reduction in cost, the improvement in speed, and what other key features and functionality that are significant to that development and what were the technological challenges associated with those, which again is easier to do than to document under the pre–FedEx standard of high threshold of innovation, being unique, novel and different from anything that’s ever been done before in software development,” said Kachinsky. “The lesser standard may make this opportunity more viable for more companies.”

But rather than advising clients to examine their previous tax refunds and the development work that they did in the past and see if they can apply for research tax credits, perhaps on an amended return, KPMG is more likely to advise them to look ahead at the software development work they are planning to do in years to come.

“I think a lot of companies are setting budgets for next year, for example, as it relates to their current IT development,” said Kachinsky. “So I think it’s a good time now to look at future planned investments from an IT perspective and understand the planned activities associated with those and planned expenditures, and then at a very bare minimum evaluate the applicability or potential for the research credits on those costs. And then, after that evaluation certainly you can look at what may apply previously, but I think the first order of business might be to look at current investments.”

“If companies have been making research credit claims for this kind of software development throughout, it’s a question of how competent they can be about whether this claim will be respected, and we think what happened in this case gives a lot of taxpayers good reasons to expect that ultimately their claims will be respected,” Culp added. “Even if the IRS continues to contest it, we think there is some good reason to think IRS Appeals and the courts will go along with what the court said in FedEx.”

But going to Tax Court to pursue a lawsuit against the IRS, as FedEx did, is probably not the best approach to take from the start. It would be better to deal with the IRS’s Appeals office if the research credit is initially rejected. “Companies that are struggling with this issue on examination might consider Appeals might be more inclined to reach a more favorable settlement,” said Kachinsky.