Both methods start from the same number
Before either method runs, you need one figure: the year’s qualified research expenses. That figure comes from activities that satisfy the four-part test and from spending that fits the statutory QRE categories - wages for qualified services, supplies consumed in research, computer and cloud rental, and the allowable share of contract research.
Everything on this page happens after that. The methods differ only in how they build the base amount - the slice of your research spending the statute treats as ordinary rather than incremental - and in the rate applied to whatever sits above it. Get the QRE number wrong and both methods are wrong; that is why a defensible study spends most of its effort upstream of the arithmetic.
One thing neither method does is decide how you deduct the same costs. That is a separate regime, and since 2025 it has its own rules - see our §174A domestic R&E expensing guide.
The Regular Credit: 20% over a fixed base
The original 1981 computation. It pays the highest statutory rate - and asks for the hardest input to produce.
- 1
Total the year’s QREs
Qualified wages, supplies, computer/cloud rental, and the allowable portion of contract research for the credit year.
§41(b)
- 2
Apply the fixed-base percentage
A historical ratio of QREs to gross receipts - for most established taxpayers the 1984-1988 base period, carried forward on the prior Form 6765. It is capped at 16% under §41(c)(3)(C).
§41(c)(3)
- 3
Multiply by average annual gross receipts
Average annual gross receipts for the four taxable years preceding the credit year. Fixed-base percentage × that average = the computed base amount.
§41(c)(1)(B)
- 4
Take the greater of that base or 50% of QRE
The base amount can never be less than half of the current year’s QREs. In practice this floor binds far more often than taxpayers expect - and it caps the whole method.
§41(c)(2)
- 5
Credit = 20% of QREs above the base
Twenty percent of the excess. Then the §280C(c) posture is applied, and any §41(f) controlled-group allocation.
§41(a)(1)
The fixed-base percentage is a historical fact
This is where most Regular-method claims quietly fail. The §41(c)(3) fixed-base percentage is the ratio of QREs to gross receipts over a defined historical base period - for an established taxpayer, the 1984-1988 years - carried forward from the prior Form 6765. It is not a number you derive from last year’s ratio because the old records are gone.
For newer companies, §41(c)(3)(B) supplies a start-up regime: the fixed-base percentage is a statutory 3% for the taxpayer’s first five taxable years with qualified research expenses, then phases into a computed ratio drawn from later years. That 3% floor is why some early-stage companies compute a larger credit under the Regular method than under the ASC - but it depends on a documented determination of which taxable year the first QREs arose in.
Whichever applies, the percentage is capped at 16% under §41(c)(3)(C).
How we handle it: our engine refuses to issue a Regular-method credit that rests on a fixed-base percentage nobody can support - it computes the figure, marks it a working-draft proxy, and blocks it from the deliverable until the real percentage is on record.
The Alternative Simplified Credit: 14% over half a three-year average
Added because the Regular Credit had become unusable for most modern taxpayers. It asks only for records you plausibly still have.
Total the year’s QREs
Same QRE number as the Regular method. The methods differ only in how the base is built.
§41(b)
Average QREs for the three preceding years
No gross receipts, no 1984 records, no fixed-base percentage - just three prior years of qualified spending.
§41(c)(4)(A)
Base = 50% of that average
Half the three-year average. Unlike the Regular method, this base is not floored against current-year QRE.
§41(c)(4)(A)
Credit = 14% of QREs above the base
Fourteen percent of the excess - or 6% of total QREs where the taxpayer had no QREs in any one of the three preceding years (§41(c)(4)(B)).
§41(c)(4)(A)
The 6% rate is easy to trigger
The 6% rate under §41(c)(4)(B) does not require a zero three-year average. It applies where the taxpayer had no qualified research expenses in any one of the three preceding taxable years - so a company with two years of QRE history and a blank third year computes 6% of total QREs, not 14% of an increment. First-time claimants and companies that paused research both land here.
Prior-year QREs must be computed, not remembered
The three preceding years’ QREs in an ASC computation are QREs as the statute defines them - not the R&D line in the financial statements, and not what a prior adviser happened to claim. If those years were never studied, they have to be computed on the same basis as the credit year, and that work belongs in the workpapers alongside the current year.
The same company, both methods, side by side
An illustrative hardware company with $2,000,000 of current-year QREs, three prior years of steady research spending, and a fixed-base percentage carried forward from its prior Form 6765.
| Step | Regular Credit (§41(a)(1)) | ASC (§41(c)(4)) |
|---|---|---|
| Current-year QREs | $2,000,000 | $2,000,000 |
| Base inputs | Fixed-base % 4.2% × $18,000,000 average gross receipts (4 preceding years) = $756,000 | Average QREs, 3 preceding years = $1,500,000 |
| Base amount | Greater of $756,000 and 50% of QREs ($1,000,000) → $1,000,000 | 50% of $1,500,000 → $750,000 |
| QREs above the base | $1,000,000 | $1,250,000 |
| Credit rate | 20% | 14% |
| Tentative credit | $200,000 | $175,000 |
| After the §280C(c) reduced-credit election (×79%) | $158,000 | $138,250 |
| Effective rate on QREs | 7.9% | 6.9% |
Illustrative only. Every figure above is invented to show the mechanics. Real results depend on your QRE composition, filing history, gross receipts, entity type, and controlled-group posture.
Now change one fact
Hold the current year at $2,000,000 of QREs but make the prior three years a ramp instead of a plateau - $600,000, $400,000, and $200,000. The ASC base drops to $200,000, the increment rises to $1,800,000, and the tentative ASC credit becomes $252,000. The Regular Credit does not move at all: the 50% floor still sets its base at $1,000,000, so it stays pinned at $200,000 - its statutory maximum for that QRE level.
That is the whole comparison in one line. Growth favors the ASC; plateaus and low gross receipts favor the Regular Credit. A company that switched from a hiring freeze to a hiring year can flip methods on nothing more than that.
Where the 6-10% figure comes from
The rough rule that a federal R&D credit lands around 6-10% of qualified spending follows from the full-rate ceilings once the §280C(c) reduction is applied - in the illustration, 7.9% and 6.9%. It is a typical band, not a floor: a first-time claimant on the ASC's 6% rate nets about 4.7% of QRE after §280C(c). Use the estimator for a first pass on your own numbers.
When each method tends to win
Tendencies, not rules. The only reliable answer is to compute both on your actual facts.
Regular Credit tends to win when…
- Your R&D spending is steady or growing slowly, so the 50%-of-QRE floor sets the base and the method pays its structural maximum.
- You are early enough to sit inside the §41(c)(3)(B) start-up regime, where the fixed-base percentage begins at a statutory 3% rather than a 1980s ratio.
- Gross receipts are modest relative to research spending - a research-heavy company with a small top line often computes a very low base before the floor.
- You already have the fixed-base percentage on a prior Form 6765, so the number is substantiated rather than reconstructed.
ASC tends to win when…
- QREs are growing fast: the base lags three years behind, so a spending ramp translates almost directly into credit.
- Nobody can substantiate a 1984-1988 base period - the common situation for companies formed after the mid-1980s or reorganized since.
- Gross receipts are large relative to research spending, which pushes the Regular base above the 50% floor.
- You need a method that survives examination on records you can actually produce: three years of QREs, not forty years of receipts.
Industry shapes the answer more than most people expect. A SaaS company scaling its engineering team usually looks like the ASC case; a mature manufacturer with flat process-development spending and heavy receipts often does not. See all four industry guides for the qualifying-activity patterns behind those profiles.
How the choice is actually made - and unmade
Method selection is an election with a shelf life, not a checkbox you revisit casually each spring.
The ASC election continues
Treas. Reg. §1.41-9(b)(1): an ASC election applies to the year for which it is made and all subsequent years unless revoked with the Commissioner’s consent. It is not re-chosen annually by default.
Revocation is an original-return act
Under §1.41-9(b)(3) the election may not be revoked on an amended return. Moving off the ASC is done on a timely-filed original return for the year you want to change.
Amended claims have their own rules
§1.41-9(b)(2), as amended by T.D. 9666, allows an ASC election on an amended return for an open year - but only where neither the taxpayer nor a controlled-group member already claimed a §41 credit for that year on another method.
In practice this means the method question is settled twice: once when you compute the credit, and once when your preparer looks at what was filed before. A study that hands over a single number without the elections analysis behind it leaves the second question unanswered. Every Ricerca study computes the Regular Credit and the ASC, records the §280C(c) posture explicitly rather than assuming it, and states the election language that has to appear on the return - see how a study runs and what lands in your deliverable package.
Short taxable years
A short credit year is not simply prorated at the end. The Regular base is prorated by days over 365 before the 50% floor is applied (Reg. §1.41-3(b)); the ASC modifies the three-year average instead (Reg. §1.41-9(c)(3)).
Controlled groups
Under §41(f)(1) a controlled group is treated as a single taxpayer: the credit is computed at group level and then allocated among members. Running the members separately produces a different - and wrong - answer.
Contract research and basic research
Only the allowable percentage of contract research reaches the QRE total, and §41(e) basic research payments interact with the base. Both change the input before either method runs.
State credits use their own base
Many state R&D credits borrow the federal QRE definition but define their own base period and rate. The federal method you choose does not automatically govern the state computation.
Calculation-method questions we get most
Which method will give us the bigger credit?
Why is the Regular Credit capped at 10% of our QREs?
We can’t find our 1984-1988 records. Is the Regular Credit off the table?
Is the ASC election something we re-choose every year?
Can we elect the ASC on an amended return for a prior year?
Does the §280C(c) election actually make us better off?
Can we make the §280C(c) election on an amended return?
More across every topic in the R&D credit FAQ, or start at the §41 hub.
Primary sources
Read the statute and the regulations yourself - they are shorter than the commentary.
- IRC §41 - Credit for increasing research activities (Cornell LII)
- Treas. Reg. §1.41-9 - Alternative simplified credit (Cornell LII)
- Treas. Reg. §1.280C-4 - Election for reduced credit (Cornell LII)
- IRS - About Form 6765, Credit for Increasing Research Activities
Tax law and IRS guidance change; paragraph numbering within §41(c) has moved with prior legislation. Verify current rules with a qualified professional before acting.