Lei do Bem: how to build a defensible R&D file

The Lei do Bem, Brazil's main R&D tax incentive, has a feature that sets it apart from other incentives: companies apply the benefit on their own when calculating tax, with no prior approval. Reporting to the Ministry of Science, Technology and Innovation (MCTI) comes afterwards, through the FORMP&D form, and the ministry's technical review may take time. If the declared activities are not recognized as research and development, the benefit may be challenged, with the legal additions.

So the right question is not "can we use the Lei do Bem?" but "can we show, years from now, that what we declared was R&D?". The answer depends almost entirely on how the project was documented while it was happening.

What the law and decree establish

The technological innovation incentives are in Chapter III of Law No. 11,196/2005, articles 17 to 26, regulated by Decree No. 5,798/2006. They include the possibility of an additional exclusion, when calculating taxable income under the actual profit regime (lucro real) and the CSLL social contribution base, of part of the expenditure on technological research and innovation development, with percentages that vary according to the increase in the number of researchers and patents obtained, as well as other benefits set out in the law.

Some basic requirements are often overlooked:

• The company must be taxed under the actual profit regime.
• It must remain tax compliant during the period.
• Expenditures must be recorded in a way that allows them to be identified by project.
• Information on R&D programs must be submitted to the MCTI, through FORMP&D, by July 31 of the year following the use of the benefit.

The deciding criterion: what counts as R&D

The legal definition of technological innovation refers to the design of a new product or manufacturing process, or the addition of new features or characteristics that bring incremental improvements and an actual gain in quality or productivity. The decree details the activity categories: directed basic research, applied research, experimental development, basic industrial technology, and technical support services.

In practice, the technical review relies heavily on the OECD Frascati Manual, the international reference for defining and measuring R&D. The 2015 edition states that, to qualify as R&D, an activity must meet five criteria:

Criterion	What it means	Helpful evidence
Novelty	Aims at new knowledge, at least for the company relative to the state of the art	State of the art review, patent search, comparison with existing solutions
Creativity	Based on original, non-obvious concepts and hypotheses	Record of hypotheses and alternatives considered
Uncertainty	The outcome, cost, or time cannot be reliably predicted at the start	Description of the technical barrier and risks at project start
Systematic	Follows a plan, with process and results recorded	Project plan, schedule, test reports
Transferable or reproducible	Results can be reproduced or transferred	Final technical report, specifications, test protocols

The weakest criterion in the files we review is uncertainty. Many companies describe well what they did and poorly what they did not know at the start. A project that develops a new product using only known techniques, with no technological barrier to overcome, may be an important market innovation and still not qualify as R&D.

Document as it happens

The most common mistake is reconstructing the story in June of the following year, in the rush before the FORMP&D deadline. By then the team has changed, emails are lost, and the report becomes a generic description of results. A defensible file has contemporaneous records, made during the project. In our experience, the items below address most weaknesses.

• Project opening sheet: technical objective, state of the art, technological barrier, hypotheses, and success criterion, dated.
• Experiment log: what was tested, the result, what was learned, and what changed in the plan. Failed tests are valuable evidence of uncertainty.
• Time tracking by project: researchers and technicians with recorded dedication, consistent with payroll.
• Accounting link: cost centers or internal orders that separate R&D expenditure from operating expenses.
• Review milestones: minutes of meetings where the decision was made to continue, change course, or stop.
• Closing report: result achieved, knowledge generated, and reproducibility.

Separate R&D from what is not

A good file also leaves out what does not hold up. Routine quality control, adapting a product to known standards, customer customization without a technical challenge, implementing off-the-shelf software, and market research are usually excluded. Including them weakens the whole file, because the reviewer starts to doubt the legitimate projects too.

A practical test we use with clients: if the technical team cannot describe in two sentences what the technological uncertainty was and how it was addressed, the project probably should not be included.

Recommended calendar

• January to March: define the year's candidate project portfolio and open the sheets.
• Monthly: check time tracking and accounting links.
• Quarterly: technical review of each project's records.
• Fiscal year close: calculate the benefit based on the already segregated expenditures.
• By July 31 of the following year: submit FORMP&D, with descriptions consistent with the records.

The Lei do Bem is one of the most relevant innovation support instruments in Brazil, and it is underused largely because of uncertainty. That uncertainty drops considerably when documentation stops being a year-end task and becomes part of the project routine.