Episode 27: Breaking Down the R&D Tax Credit

Transcript

IRIS LAWS

On this episode, we're focusing on the research and development tax credit. What qualifies for the incentives? How does it work? Is it risky? We welcome Mike Boenzi and Will Stallings, leaders on our Federal Specialty and R&D tax team, to break this all down and more. From your one stop for tax updates and analysis, I'm Iris.

DEVIN TENNEY

And I'm Devin.

IRIS LAWS

It's Tuesday, June 2, and this is “Tackling Tax.”

Well, believe it or not when I first started at the firm, I had a brief stint helping with a few R&D credit projects, which frankly, was enough for me to know that this area can get super complicated. So, I'm hoping my two guests today can help dispel some of my PTSD from what I remember, being fresh out of college and being super overwhelmed by everything that goes into these studies.

So thankfully, we have two of the most experienced folks from our firm on the call to help us through this. We have Will Stallings and Mike Boenzi. Mike is a partner with the firm who's been serving clients in this space for almost 25 years, and Will is a part of the same team and also helps with M&A and other projects in our Federal Tax Specialty Group.

So, welcome to “Tackling Tax.”

MIKE BOENZI

Thanks, Iris. Great to be here.

WILL STALLINGS

Yeah, thanks for having us.

IRIS LAWS

So, on its face, right, I would guess most of our listeners have some, maybe, like mental visual of what they think R&D is so, right? Lab coats, beakers, that kind of thing. Will, I'll start with you though. What actually is considered to be R&D for this purpose?

WILL STALLINGS

Yeah, it's a great question. So, for tax purposes, the definition of R&D is much broader than just the white lab coats and beakers—R&D in the traditional sense—that a lot of us think about. It's really about solving technical problems that have uncertainty with them while also developing anything that's new or improved from a functional standpoint.

So, to kind of put that into terms that the Congress and the IRS give us, that would be what they call the four-part test, where you have to be working on developing something that's functionally new or improved; you've got to have some level of technical uncertainty with what you're doing; you have to be relying on hard sciences, so, engineering, biology, chemistry, similar type disciplines; and substantially all of the work you're doing must involve a process of experimentation.

So, those are really the key attributes from a tax standpoint that I would say comprise the definition of R&D. So, qualifying R&D, it shows up in some places that are somewhat unexpected sometimes. It's not just making, let's say, the next new widget. It could be something that's either tangible or intangible. So, for example, a lot of our clients in the manufacturing space work on developing process improvements, whether it's relaying out their machinery and equipment for a manufacturing line, that can qualify for this tax incentive.

If you're working on building custom software solutions, whether it's iterative developments or upgrades or enhancements, versus just an entirely new software platform, that can qualify. Food and beverage companies with their formulation and their overall techniques they use in developing products that they're going to be selling, that can qualify. Engineering-based service solutions, construction, architectural engineering, those industries may also qualify.

You know, the overall, I would say, breadth of what can and can't qualify; there's not really any industry that we just say, yeah, there's no chance there's going to be R&D. And I would say one key component of that is the fact that so many companies are investing in software development and AI these days, that that kind of opens the door for any industry that's investing in that. So, you potentially qualify for claiming a credit.

So, things like if you're in the retail industry or banks, financial institutions, insurance companies, healthcare, those can also qualify. We have plenty of clients in that space. And one final comment is that this credit it's not really a revolutionary credit. This is often an evolutionary credit. So, if you're making incremental improvements to products or processes or software, that can very much qualify, you don't have to be creating something brand new to the world.

DEVIN TENNEY

Will, from what I'm hearing, is there's a lot of different taxpayers and companies out there that may be doing R&D and maybe qualifying for this credit, but really, what is the credit? What is something what's important about this? Why should our listeners really care about what they're doing with their R&D and what might that opportunity be?

WILL STALLINGS

Yeah. So, this is a dollar-for-dollar tax credit. It's not just a tax deduction, which that's a separate but related issue. But the credit itself is really based upon a calculation that is composed of up to four qualifying types of expenses. So, wages, supplies and materials, payments to contractors, and then cloud computing or hosting costs called computer rental costs.

Those costs, when they're eligible to count towards the credit, then compute the credit, which is often for federal purposes, about six to eight cents on every dollar of qualifying R&D spend. From a state standpoint, that can be roughly three to five cents on the dollar, but states vary. Roughly 38 of our 50 states do have some level of R&D credits, so it's a nice benefit on top of the federal credit. The credit is obviously going to be reducing your income tax liability.

That being said, for certain startup companies that meet certain criteria, they can also get refunds for the employer portion of their Social Security, FICA, and Medicare taxes as a refund if there's not sufficient tax liability in the current year. Furthermore, because it's a Section 38 general business credit, you can carry the credit back one year and forward up to 20 years. If you ever have unused credits sitting inside of a corporation, for example, and you go through a stock sale, those may carry over to a buyer. Now, there's certain tax attribute limitation rules to consider, but there is more than one way to use this credit.

Short answer is that, you know, roughly on average, 10 to 12 cents on the dollar between federal and state, if state credits are available, is what you're going to be looking at. It's a great way to help subsidize the cost of your development activities, as you're seeking to create new products for the marketplace.

IRIS LAWS

Well, I mean, that's substantive, right? So, Mike, if we shift a little bit to you, if someone's, you know, going to go through this process, they still might have some misconceptions. So, what are those misconceptions that you might hear on a regular basis from your clients?

MIKE BOENZI

I would say, you know, probably the biggest misconception that I hear quite frequently is that, and you kind of alluded to this and Will kind of addressed this in some of his comments a few minutes ago, but that the credit is only applicable to companies with folks in white lab coats, or requires new-to-the-world type of scientific breakthroughs.

And in reality, the credit is really an incentive that is meant to reward evolutionary types of development. I would say another misconception that I commonly hear is that a company has to be in a traditional, quote unquote, “innovation industry” like tech or life sciences. And the reality is that’s simply not true. Manufacturers, construction companies, food and beverage businesses, engineering firms, it really runs the gamut across virtually every industry in terms of the potential for qualifying activities and credit.

And in fact, many of our listeners might be surprised to to learn that a significant portion of the businesses that claim the credit on an annual basis are actually in the manufacturing space. In addition to those couple thoughts, we also hear folks say, well, our project succeeded so it probably doesn't qualify for the credit. And that is simply not true.

The credit is not limited to simply failed projects. Success or failure is really irrelevant from the standpoint of qualifying for credit. On the other side, some folks think that every single technical project automatically qualifies for the credit, and that's not, unfortunately, correct either. As Will had previously mentioned, we still have to show that technical uncertainty and a systematic trial-and-error process capable of evaluating more than one alternative exists.

In addition to that, there are some misconceptions out there around documentation. And that is, if I don't have perfect documentation, or for example, as a company, we don't have a formal time tracking process in place, we can't claim the credit. And the reality is perfect documentation is certainly not the standard, and the time tracking is not required to claim the credit.

That said, you know, given the burden of proof rests with taxpayers in substantiating their entitlement to the credit, it certainly important to work with a trusted advisor that really has their finger on the pulse of the key issues in today's exam environment and is really experienced in leveraging taxpayer documentation, in whatever form it may exist, to help tell the story in full view of recent court decisions and current expectations of the IRS based on today's exam environment.

Some other things I hear; well, you know, as a business, we have to be profitable to benefit from the credit, or spending has to jump year over year to qualify, or only large corporations can qualify for the credit. The reality is, the credit is applicable to not only startups, not only middle market firms, but also large multinational companies across virtually every industry.

Even if recurring spend remains level, and that's by virtue of the simplifying credit which came into existence back in 2006. Now, startups and early-stage businesses potentially have a unique opportunity to convert what ordinarily is an income tax credit into a payroll tax credit. And what's more, a business or its shareholders or partners that cannot fully utilize the credit in the year it's generated due to its taxable income situation, any unused federal credits can first be carried back one year and then carried forward up to 20 years to offset future tax liability.

A lot of clients also assume that claiming the credit will automatically mean that you're going to be examined. And while it's true that over its 45-year history, the credit has enjoyed a fair amount of scrutiny by the IRS and has become a coordinated issue that has required centralized risk assessment within the IRS, it doesn't mean companies should have avoid claiming the credit if they have a legitimate claim and worked to compile good supporting documentation.

The reality is a very small percentage of tax returns that claim the credit on an annual basis are examined. And then finally, one last comment and thought that I'll offer is, you know, many companies think, well, hey, we looked at this years ago and didn't qualify, so that means it still must be true. And that couldn't be further from the truth.

Certainly, businesses evolve, projects change, and new court cases expand the scope and activities of costs that potentially qualify for the credit. And the facts may look very different now. Given the shifting and evolving nature of the R&D credit landscape, I think it's best practice to periodically revisit potential opportunities to claim the credit.

DEVIN TENNEY

Yeah, that's a great point, Mike and I want to dig in a little bit about what those qualifying activities and costs actually are. I'm assuming there's probably some that do and don't qualify that surprise people. You know, in practice, where are companies really, you know, seeing opportunities where maybe they aren't claiming these credits and they probably should be or potentially they're trying to overclaim the credit?

MIKE BOENZI

Well, I would say first and foremost, it's certainly important to keep in mind that businesses don't need to be developing things that are revolutionary to the industry in order to qualify. And we talked a little bit about how this is an evolutionary type of incentive. I would say the areas where I commonly see companies are potentially underclaiming or missing out on tax savings opportunities.

First and foremost are process development and improvement. That is often an area of overlooked opportunity. In addition to that, you know, when thinking about the credit, a lot of businesses have a tendency to focus on low-hanging fruit, that is, employee wages for direct research, and miss out on things like direct supervision and support wages, or potentially prototype and pilot model supply costs, or even cloud computing expenses, as well as contractor costs when there is a perceived funding situation involved.

State credits, Will kind of touched on this a little bit in his commentary, you know, some taxpayers only focus on the federal side when the reality is nearly 40 or so states have some form of R&D incentive on the books, many of which are refundable, in fact, and offset income, franchise, or sales and use, gross receipts, or payroll taxes.

Another area of opportunity that I commonly see is, is businesses often fail to review prior open tax years under the statute of limitations for potential amended return claims. In addition to that, I would say situations involving funding, let's say, between a business that's performing research and development activity and potentially a customer, are situations that offer that often are ripe for opportunity.

And just because an item is sold to a customer or a customer may be paying for a product where development work is occurring, you know, that doesn't automatically exclude those activities and costs from the credit. It really depends on, at the end of the day, which party in that two-party relationship bears economic or financial risk and which party retains the underlying rights to the research effort.

And then, still another area of opportunity I commonly see is global R&D incentives, in fact. So, for multinational businesses that are performing R&D outside of the U.S., most industrialized countries of the world have some form of R&D incentive on the books and are continuing to become more and more aggressive just in terms of their efforts to attract intellectual property.

Taking a global view can oftentimes uncover opportunities not previously considered to lower a company's global effective tax rate. Now, on the flip side, we talked a little bit about areas of opportunity. On the flip side, companies can potentially overreach when they sweep in activities that are really, I would say, more routine or commercial in nature. For example, routine testing that's part of everyday common production.

Or another example might be purely cosmetic or stylistic changes to a product or adapting an existing product that doesn't involve technical uncertainty, or maybe routine data entry or administrative work. Another area where companies can potentially overreach is software, especially where the work is more standard implementation or option selection rather than actual development involving a systematic process of experimentation.

You know, while the IRS or state exam fears are often overstated, I would say documentation of the underlying facts coupled with contemporaneous artifacts are key to substantiating a claim for credit. There seems to be, I would say, a handful of recurring lessons from recent court cases. The first is documentation evidencing real experimentation. That is, evaluating alternatives to resolve uncertainty is important.

In addition to that, a critical view of funding arrangements to ensure we don't run afoul of the funded research exclusion is important. The use of defensible estimation methods supported by contemporaneous records, that seems to be a common theme, and establishing nexus between employees, projects, activities, and costs are also important in being able to properly substantiate the credit.

IRIS LAWS

Well, that's all hugely helpful guidance. I do want to kind of circle back a little bit to a very specific one that you mentioned, which is software, just because I know that that was sort of a hot button legislatively, you know. Guidance came out not too long ago, right, about that. So, Will, do you want to dig in a little bit more on software specifically because that might be an opportunity for maybe some less traditionally like, some industries that you might not traditionally think of could qualify in that area. So, any quirks we should be aware of?

WILL STALLINGS

Yeah, absolutely. So, with software, I would say kind of the key thing to think about is, well, I guess there's two key things. One is going to be what are the types of software related activities that are being performed? And two, what is the intent or purpose of that software?

So, for example, with the activities, in 2023, a ref product, or an IRS notice rather, came out that helped provide some clarification around what activities may or may not qualify for Section 174 treatment, which is often very synonymous with Section 41 treatment. Not the same, but similar principles, right? And when you're looking at the activities being performed to qualify for the credit, you still have to have uncertainty, but it can't be something like training or maintenance or basic installation implementation.

Now the key aspect, though, is going to be the intent or purpose of that software that's being developed. So, what the IRS wanted to weed out would be any internal use software. It can still qualify, but you have to meet these three additional tests. Now for tax credit purposes, R&D credit purposes, internal use software is defined as software that is basically intended to serve a general or administrative function or purpose.

So, things like automating your accounts payable system, working on CRM automations internally only by the taxpayer's use, for example. Those can still qualify, but you have to have these three additional tests met in addition to the four-part test that we mentioned at the beginning of the podcast. These three additional tests are; that the software is essentially highly innovative, so kind of a step above and beyond just new or improved functionality; you've got to have significant economic risk with the overall software development effort; and it can't be something that's commercially available without having to do substantial modifications.

So, those are the three additional tests. Now, if you're doing, let's say, customer-facing software or software that's not intended to be for a general or administrative purpose, that gets you out of those additional three tests, and you can just go back to documenting the traditional four-part test. So, there is a little bit of IRS guidance on the software component, but not a lot of detail within the Treasury regs or the code other than that distinction you just mentioned.

IRIS LAWS

Well, and along those lines, I know we mentioned, you know, the hot button that everyone's talking about, which is AI, a little bit earlier. But, Mike, if we switch to you in that same vein, what about AI, right? Is it considered software?

MIKE BOENZI

Generally, I would say yes. AI work is generally considered a form of software development but certainly depends on the facts and circumstances of the project, not just the label, quote unquote, “AI” itself. What we're seeing is clearly AI is disrupting businesses by shifting, you know, from a tool of incremental efficiency to really a core driver of strategy that's accelerating automation and fundamentally altering competitive advantages across industries.

And as we're talking with our clients, AI is really enabling businesses to move, I would say, from a one-size-fits-all type of approach to hyper-personalization, even tailoring services to individual customer behaviors at unprecedented speed. And in fact, a lot of companies are using AI to accelerate innovation themselves and launch products faster with less risk, shifting from long development cycles to continuous, agile type of innovation cycles.

And as we're having conversations with clients, we're hearing about instances where clients are either evaluating the potential for AI applications in their business or currently making—or plan to make—innovation investments leveraging AI. And depending on the client's industry, these investments can be significant.

Now, in terms of the alignment between development involving AI and the research credit. In many cases, the work that's being done by organizations with respect to AI involves solving complex technical problems and is often focused on creating things like new or improved products or processes, or enhancing data-driven decision making, developing agents or models, or optimizing workflows, right? And AI development, I would say, frequently involves technical uncertainty at the outset. It also involves a systematic experimentation process involving forming an initial concept, and testing, and refinement. And oftentimes those efforts are grounded in computer science, data science, engineering. And the reality is, when these characteristics are present, AI initiatives can represent a meaningful opportunity to offset the cost of innovation through the credit.

And when it comes to AI, there are really thousands upon thousands of potential use cases, and businesses across virtually every industry are leveraging both generative and agentic AI in a variety of different ways. Given the variety of potential use cases for AI, I would submit that there is a “spectrum,” quote unquote, of activities, some of which may qualify for the credit and others which likely will not.

For example, on the one end of the spectrum, you could have businesses that may be using, quote unquote, “off-the-shelf” AI tools like OpenAI's ChatGPT or Anthropic’s Claude to draft an email or summarize notes from a meeting or format data in a certain way, or simply integrate a purchased AI tool. With this type of AI use, testing is often limited to evaluating the effectiveness of configuration rather than new features, new functionality that are developed aside from the core functionality of the off-the-shelf tool.

On the opposite end of the spectrum, I would say businesses could be leveraging AI to develop things like new algorithms or novel model architectures, or custom models, or predictive systems, or even AI-enabled products. Those particular use cases often will exhibit the requisite technical uncertainty and experimentation and thus have a higher likelihood of qualifying.

And even then, you could still have scenarios in between those two extremes where businesses could be using AI to draft code, as an example, or build apps, or develop prompts or agents. In these instances, qualification is going to really depend on the systematic trial and error process that's undertaken.

The evaluation, really, of R&D credits for AI differs from generative AI to agenetic AI, in that generative AI uses large language models to create content through prompts, whereas agentic AI involves building autonomous systems that have the capability to plan and execute multi-step workflows and even have the ability to adapt to their environments, typically involving a more intensive process of experimentation.

Generative AI requires documentation showing how the AI model was modified or trained on proprietary data, rather than just used, whereas agentic AI, I would say, requires documentation that evidences the development of custom software components, or things like AI/API integrations or decision-making logic. So, the bottom line when it comes to evaluating AI in the R&D credit is like many other things in the R&D credit space. It really, truly depends on the facts and circumstances.

IRIS LAWS

Well, you answered my question. I was going to ask about, well, what about developing an agent? Because I know we've been playing with some of that, so that is right on the nose and super helpful. So, you know, stepping back, we've talked about what R&D is, what the opportunity is and things to watch out for all of that. But Will, what about from a strategic perspective, right? When when should someone actually claim the credit and how does it interact with some of your other tax planning possibly?

WILL STALLINGS

Yeah. So, the timing does matter for when you claim an R&D credit. There are various things to consider. One of those being, you know, how you're treating the cost for tax purposes. So, under Section 174, for example, are you liking to deduct those R&D costs immediately or defer them over five years if they're domestic. When you're claiming credits, if you ever decide to sell the company, then that could have a big impact, because if you have any unused credits, or credits that have been used in tax diligence from a buyer standpoint, they're going to scrutinize those credits pretty hard and want to make sure that they're legitimate claims with the right documentation in place.

Making a 280C election to tax effect, the credit is important. So, you can't do that at all if you amend the return and want to basically make an election where you didn't previously make one. Claiming a credit when you amend the return, as Mike mentioned earlier, that has implications, because if you say, oh, we'll just claim the credit later when we amend.

Well, while you can do that, if the statute is open for those years, that's going to trigger an automatic review by the IRS to process that return, which requires documentation to be submitted when you amend. On top of all of this, R&D studies are very much time-intensive engagements. These are not things that you do at the 11th hour, not quick exercises.

There's a quantitative and equally as important qualitative aspect to these projects. So, you want to make sure that you have enough time before the tax prep team needs to incorporate this into a tax filing to make sure that it's vetted, it's calculated correctly, and you're claiming it on the right federal and state return. One last point with the timing is that this relies not just on the current year R&D data, but it relies on at least the prior three years R&D data if you're doing a certain credit calculation methodology.

But the short answer is that it does depend on historical R&D spend by the company, including spend from any acquisitions or removing it from dispositions. But it's going to be more than just the current year is the short answer. A best practice for tax leaders and CFOs would be make sure that we're treating R&D credits as a proactive item, not as an afterthought. If it's an afterthought, we're probably going to be leaving dollars on the table at the end of the day.

And then finally, the credits should be considered in connection with certain other tax-related items. So, just mentioned Section 174, the tax deductibility of a lot of those costs that also help generate a credit, whether or not you make a Section 280C election to tax effect the credit. There's various international tax implications such as EMP calcs, foreign tax credit computations. And then if there ever is a change in ownership, Section 382 for corporations could apply, things like Pillar Two.

So, there's a multitude of different tax provisions that would also apply or else be related in some form or fashion to claiming a federal R&D credit.

DEVIN TENNEY

Well, Mike, if we zoom out a bit, clearly one of our favorite topics on the podcast is tariffs. We've talked about its impact on construction and real estate and the energy markets. How is it influencing R&D?

MIKE BOENZI

Great question. And certainly, that is a very hot topic with a lot of our clients today. In my opinion, when it comes to tariffs, I don't know that we necessarily have a broad cross-section of knowledge with which to draw a definitive inference given how recent the imposition of tariffs has been.

On the one hand, we have some economists that have maintained that tariffs usher in things like increased prices, or inflation, or reduced productivity, or a decrease in innovation due to the lack of foreign competition. While others assert tariffs will bolster the manufacturing sector, as an example, by essentially levying a tax on imports.

It seems there may be two schools of thought here. Ultimately, on the one hand, tariffs may be negatively influencing investment in R&D over the long run, primarily because they raise input costs, or create uncertainty, or shrink overall market access. But still, on the other hand, many economists, including the Economic Policy Institute, have suggested that tariffs may actually boost profits for specific sectors or industries, potentially creating additional capital to invest in things like R&D.

In addition, higher costs and supply chain disruptions are forcing product redesigns or material substitutions and product and process re-engineering, helping to positively influence investment in R&D. What I think is not in doubt is the incentive effect of the credit as a matter of policy. While tariffs may increase the cost of doing business, the R&D credit subsidizes production and innovation through the reduction of tax liability. And in light of this, the R&D credit can actually be used as a mechanism to offset the cost of tariffs and help to provide a boost to innovation domestically.

In addition to tariffs, and this is something that Will alluded to a few minutes ago, I think the requirement that continues to remain in place following the One Big Beautiful Bill Act to capitalize and amortize foreign Section 174 costs over a 15-year period, as well as some of the import duties that continue to exist, is causing some companies that perform R&D offshore to consider potentially reshoring some or all of those activities and in turn, expenditures, the cost of which could be offset through the credit.

That said, I think the ability for companies to pivot their supply chains and business operations is not always necessarily quick, nor is it practical in all instances. Also, the specific impact to businesses can certainly vary depending on their facts and circumstances.

IRIS LAWS

Well, what a comprehensive discussion. Thank you guys so much for coming. We really enjoyed it and we look to have you back on the the podcast soon.

MIKE BOENZI

Sounds good. Thanks for having us, Iris.

WILL STALLINGS

Yeah. Thank you so much for having us.

DEVIN TENNEY

Each episode will bring you what we call a “Focused FORsight of the week,” an article or webinar that might be of interest to you. Going along with our last topic of conversation today, there is an article titled “Opportunity Knocks: R&D Credits Amid Tariffs,” authored by one of the leaders of our R&D team that dives deeper into this interesting situation. Take a look for some interesting visuals and discussions on some broader trends in this space.

IRIS LAWS

And that's our show. Thanks for joining. Remember to subscribe and listen in for the next episode of the podcast. Until next time.

ANNOUNCER

The information set forth in this podcast contains the analysis and conclusions of the panelists based upon his, her, or their research and analysis of industry information and legal authorities. Such analysis and conclusions should not be deemed opinions or conclusions by Forvis Mazars or the panelists as to any individual situation, as situations are fact specific. The listener should perform their own analysis and form their own conclusions regarding any specific situation.

Further, the panelists’ conclusions may be revised without notice, with or without changes in industry information and legal authorities.

Related FORsights