US GAAP expenses virtually all R&D as incurred. IFRS/Ind AS capitalize only qualifying development. Economically, a pharma company's molecule pipeline and a software firm's codebase are capital, built by spending that was expensed. So for research-heavy firms both halves of the ratio are distorted, usually in offsetting directions. The numerator has EBIT hit by this year's R&D, and the denominator has no research asset. Damodaran's standard repair runs in four steps:
- Choose an amortizable life matching how long research keeps paying off: ~3–5 years for software, ~10 for pharma. His R&D converter spreadsheet ships with suggested lives by industry.
- Build the research asset = this year's R&D × 100% + last year's × (n−1)/n + … (straight-line fade over life n).
- Adjust EBIT: Adjusted EBIT = EBIT + this year's R&D − amortization of the research asset. The after-tax fix adds the full difference (R&D − amortization) to NOPAT, with no extra tax effect, because the tax benefit of expensing was real cash and stays.
- Adjust IC: Adjusted IC = IC + unamortized research asset.
Advertisement
How does capitalizing R&D change Microsoft's ROIC?
Here's a mini-build on Microsoft (illustrative, modeled on ~FY24/FY25 scale; pull the actual 10-K; $bn). R&D spend over the most recent five years runs 25, 27, 29, and this year 32. Take a 3-year life. Research asset = 32 + 29 × ⅔ + 27 × ⅓ = 32 + 19.3 + 9.0 ≈ 60. Amortization this year = (29 + 27 + 25) ÷ 3 = 27. Adjusted EBIT = 128 + 32 − 27 = 133. Adjusted NOPAT = reported NOPAT + (32 − 27) = 105 + 5 = 110. Adjusted IC = 320 + 60 = 380. (You'll build the 105 and 320 in Worked Example 1.) ROIC moves from 32.8% to 28.9%. That's still elite, and it's closer to the truth.
Part of Microsoft's "capital-light magic" is capital the accounting never booked.
Carry the general pattern around with you. For firms with growing R&D the adjustment lowers measured ROIC, because the asset builds faster than the amortization catches up. For steady-state R&D it barely moves, and Practice 7 proves this. Either way it can rerank peers, say a heavy self-investor against an acquirer of finished technology. The same logic extends to brand-building advertising for consumer companies (Mauboussin's intangibles work). Measured ROIC in the intangible economy systematically overstates asset-light champions. Adjust when you compare across R&D intensities, and skip it when you compare like with like.