Before You Claim a Puerto Rico R&D Credit, Read This

Most of the trouble we see with research credit claims does not come from aggressive positions. It comes from ordinary companies doing real research and handling the claim badly, usually in ways that were easy to avoid and expensive to fix.

This is the article we wish companies read in year one instead of year three.

The framework, stated once and plainly

Puerto Rico's current consolidated incentives statute is Act 60-2019. It replaced the earlier collection of laws, including Act 73-2008 (Legacy), which still governs some grandfathered decrees on the island but is not the basis for new planning.

Research and development activity is addressed under §3030.01. The model is two fixed rates. Fifty percent applies to direct R&D investment deployed in Puerto Rico into experimental activities. Twenty percent applies to indirect R&D investment deployed outside Puerto Rico or into incidental activities.

Note what is not in that description. There is no negotiable rate. There is no qualification percentage to argue about. There is no base period calculation. This is Puerto Rico law, not the US federal research credit, and importing federal mental models into the analysis causes confusion rather than clarity.

Because the rates are fixed, every dollar of value comes down to two determinations: is the activity experimental, and where was it carried out.

Mistake one: building the claim backwards

The most common failure pattern is simple. The year ends, someone decides to pursue a credit, and the work of reconstructing what happened begins.

Engineers are asked to remember what they worked on fourteen months ago. Time allocations are estimated from memory. Projects are defined after the fact, around the costs rather than around the technical objectives. The resulting file is a reconstruction, and it reads like one.

A reviewer can tell the difference immediately between documentation created while work was happening and documentation created to support a filing. The first describes a problem and an investigation. The second describes a conclusion and works backward.

The fix is unglamorous and cheap: define development projects at the start with a stated technical objective and the uncertainty being addressed, capture time against them as the work happens, and write a short closing summary while the details are still fresh. Three paragraphs written in the moment are worth more than thirty pages assembled later.

Mistake two: claiming everything

There is a temptation, particularly when the rate is attractive, to sweep in as much cost as possible and let the claim get sorted out later if anyone asks.

This is counterproductive even on purely selfish terms. A claim that includes obvious production costs, routine quality control, and ordinary maintenance invites a reviewer to question the whole file, including the parts that were genuinely excellent. You spend the examination defending the indefensible items instead of explaining the real research, and the real research gets less sympathetic treatment as a result.

Clean claims get clean outcomes. Padded claims get scrutinized in full.

Be particularly careful with the geographic rule. Work performed outside Puerto Rico does not become island research because the island entity paid for it. That is precisely the situation the indirect category addresses, at a different rate, by design.

Mistake three: ignoring the uncertainty test

If you take one analytical idea from this article, take this one.

The question that determines qualification is whether there was genuine technical uncertainty at the outset, and whether your team resolved it through systematic investigation.

Not whether the work was difficult. Not whether it was performed by scientists. Not whether it happened in a lab. Not whether it was important to the business. Uncertainty and systematic investigation.

Running a validated process is not uncertain. Executing an established test method is not uncertain. Installing equipment per the vendor manual is not uncertain. Implementing a known software pattern is not uncertain. All of these can be demanding, valuable, and expensive without being research.

Conversely, work that looks mundane can qualify. A batch investigation where nobody knew the mechanism. A process transfer that failed and required real experimentation to fix. A tooling problem where the approach had to be invented. These are often filed as operations and are frequently legitimate research.

Mistake four: treating the decree as background

Your benefits come from a decree. A decree is a negotiated document with specific terms, conditions, employment commitments, and reporting obligations attached to it.

Companies regularly plan against a general understanding of the framework while never carefully reading their own decree, and then discover a condition they have not been satisfying. Conditions can relate to employment levels, investment commitments, the scope of covered activity, and reporting deadlines. They are not decorative.

Read your decree. Make sure someone owns compliance with its specific terms. If your operations have changed materially since it was issued, find out whether that matters before someone else does.

Mistake five: no technical narrative

Financial substantiation without technical substantiation is half a claim.

Someone has to be able to explain, to a technically literate reviewer who was not in the room, what the project was trying to achieve, what was not known at the start, what approaches were attempted, which ones failed, and how the uncertainty was eventually resolved.

Companies with excellent cost tracking and no technical story struggle badly in examinations. Companies with a clear technical story and reasonable cost allocation generally do fine.

Life sciences and aerospace companies have an advantage here that they rarely exploit. Your quality system already contains protocols, reports, deviations, and change records that document exactly this. The material exists. It was simply never organized for this purpose.

Mistake six: no owner

A research credit claim touches finance, engineering, operations, and often quality. When it belongs to everyone, it belongs to no one.

The pattern that works is a named owner with authority to ask engineering leaders uncomfortable questions and to say no to costs that do not belong. The pattern that fails is a year-end exercise coordinated by whoever has capacity in Q1.

What a healthy position looks like

Development projects are defined before the work starts, with a technical objective and an articulated uncertainty.

Time is captured in a way that separates development from production support, by people who understand the distinction and apply it consistently.

Costs tie to actual accounting systems rather than retroactive estimates.

Each closed project has a short technical summary written near the time, describing the problem, the approach, the failures, and the resolution.

Activity performed outside Puerto Rico is classified as such without argument.

Someone reads the decree and owns compliance with its terms.

None of this is complicated. It is just work that has to happen while the research is happening, which is precisely why it so often does not.

The disclaimer, and it is a real one

Everything in this article is indicative advisory context. Qualification, rates, decree terms, filing mechanics, and administrative practice depend on your specific facts and change over time. Nothing here is a determination about your situation, and you should confirm your position for your own circumstances before relying on it for planning.

The encouraging part

Having said all of that, the honest overall message is positive.

The Puerto Rico research credit is a serious, well-defined incentive for companies doing genuine development work on the island. The rates are clear. The rules are workable. The standard is one that real research organizations meet comfortably.

Companies that do real work and keep an honest record do well with it. The ones who struggle are almost always the ones who treated substantiation as an afterthought, and that is a solvable problem if you start now rather than next year.