# Research tax incentives and the composition of business spending

Policy Brief · United States · Innovation and Productivity

Research tax incentives can change the amount, timing, location, and reported classification of business spending. Those responses should be distinguished before a rise in tax-eligible expenditure is interpreted as an increase in innovation. The tax definition of qualifying activity is necessary for administration, but it is not itself a measure of new knowledge or commercial productivity.

The U.S. framework makes the distinction especially important because the deduction for research expenditure and the research credit are separate provisions. Their eligibility, timing, and interaction affect the benefit received by a business.

## The deduction and the credit differ

The IRS's [Revenue Procedure 2026-32](https://www.irs.gov/irb/2026-39_irb) describes accounting-method procedures following the 2025 addition of section 174A. It explains the treatment of domestic research or experimental expenditure for taxable years beginning after December 31, 2024, separately from the continuing amortization treatment of foreign research under section 174. Historical costs and transition elections require their own analysis.

The [December 2025 instructions for Form 6765](https://www.irs.gov/instructions/i6765) describe the research credit and its qualified-research requirements, including the business-component and experimentation tests. An expenditure treated as research for deduction purposes does not automatically establish an identical credit claim. The instructions also distinguish categories of in-house and contract research expenses.

These sources provide an institutional starting point. They do not establish that every expenditure labeled research in a company's financial statements qualifies under either provision. Nor does the accounting treatment determine whether the expenditure would have occurred without an incentive.

## Timing affects the financing of research

An immediate deduction can reduce the tax cost of current expenditure sooner than amortization over several years. The economic value depends on the taxpayer's ability to use it, the relevant tax rates, and the alternative timing. A company with losses may experience a different cash-flow benefit from a profitable company, even if both undertake the same research.

The distinction matters for financing-constrained businesses. Earlier tax relief can preserve cash during a project whose revenue is uncertain or distant. For a business with ample funding, the same relief may primarily change the after-tax return. These are different channels and should not be assumed to generate the same spending response.

A change in the treatment of prior expenditure can also improve cash flow without financing a new research activity in the same year. An evaluation should separate relief associated with historical costs from the incentive facing the next project. Otherwise, a large tax benefit can be incorrectly matched to a current research increase.

## Eligible inputs are not innovation outputs

| Observation | Economic information | Additional evidence required |
| --- | --- | --- |
| Qualified research expense | Spending within the applicable tax definition | Whether total research resources increased |
| Research payroll | Labor allocated to specified activities | Hours, skills, and changes in compensation |
| Contract research | Purchased work under relevant arrangements | Who performs the work and bears the risk |
| Patent or technical result | An observable research output | Quality, use, and additionality |
| Productivity improvement | Output relative to measured inputs | Attribution to the supported research |

The distinction between prices and quantities is important. Research expenditure can rise because specialist wages increase while the number of research hours remains unchanged. Higher spending may then reflect competition for scarce skills rather than a proportional increase in experimentation. Staffing, hours, and project evidence help interpret the financial total.

Outputs also arrive with a lag and are uncertain. A technically unsuccessful project can still generate useful knowledge. A patent can have little commercial value, while a process improvement can be valuable without being patented. A single output count is therefore an incomplete measure of the economic return.

## Composition can respond before total spending

A tax incentive can alter which projects are selected. A business may favor activities whose costs fit the tax definition, shift research location, change the balance between employees and contractors, or accelerate expenditure. These responses can be real without representing an equivalent increase in total research over a longer period.

Some reported growth can also result from improved documentation or a reassessment of costs. That is an administrative response rather than evidence of additional technical work. It need not be improper: a business may identify qualifying activities more accurately. Evaluation should nevertheless distinguish reclassification from a change in resources devoted to research.

The appropriate counterfactual depends on the policy objective. If the aim is to retain domestic activity, location matters. If the aim is to raise total experimentation, a relocation from another jurisdiction is not the same result as a new project. If the aim is knowledge spillovers, the diffusion of results beyond the recipient becomes relevant.

## The tax base and the business organization interact

Businesses differ in their ability to claim and use incentives. Group structure, contractual rights, the allocation of expenses, and other tax provisions can change the net benefit. Comparing headline statutory percentages across firms or countries therefore risks treating different effective incentives as identical.

Administrative costs matter as well. Documentation can improve accountability while creating a fixed cost that is more burdensome for small firms. A policy assessment should consider whether the requirement helps establish genuine qualifying activity and whether its implementation creates avoidable barriers. Simpler administration is valuable when it preserves the evidence needed for a reliable claim.

## Assessment

Research incentives should be evaluated through the resources and decisions they change. The deduction and credit frameworks establish the tax treatment; additionality requires a separate comparison with activity that would otherwise occur. Timing relief, location changes, wage pressure, and reclassification should remain visible in that comparison.

For business leaders, the relevant issue is the net incentive facing a documented project. For policy, it is whether public revenue supports additional useful research and wider benefits. The interpretation of the fiscal estimate is examined in [Tax expenditure estimates and the cost of investment incentives](/research/tax-expenditure-estimates-and-the-cost-of-investment-incentives/). This brief is economic analysis rather than a determination of an individual taxpayer's eligibility.

---

American Economic Senate