A rate qualifies as risk-free for your valuation only if two conditions hold:
- No default risk. The promised cash flow arrives with certainty.
- No reinvestment risk at your horizon. You're valuing decade-scale cash flows, so you want a long rate, the 10-year government bond. A T-bill would force you to guess reinvestment rates nine more times. It's the same mismatch trap you met when building the ERP, and it applies here in the same way.
How does currency decide the risk-free rate?
The currency rule. Learn this one as law. The currency of the cash flows dictates the currency of the risk-free rate, and of every rate stacked on it. If you forecast in rupees, you use an INR risk-free rate and INR-consistent premiums, which give you an INR WACC. If you forecast in dollars, the whole stack is in USD. Headquarters, listing venue and reporting currency don't matter at all. Only the forecast's currency does. Mix them, say with a USD Rf under INR cash flows, and you smuggle a ~2-point inflation differential into your rate and revalue the company by a third without noticing. Done consistently, an INR valuation and a USD valuation of the same company should agree once you convert at Fisher-consistent expected exchange rates. Changing currency changes the units, never the value.
In the US it's easy, because the 10-year US Treasury is the standard USD risk-free rate. Its market is deep and liquid, and its default risk is treated as nil. A printing press backs the promise. The risk you carry is debasement, and that gets priced as inflation. Pull it from FRED, series DGS10. Say 4.2%, the illustrative mid-2026 value from the risk-and-return work. That's Station 1 done, in one line.
Advertisement
Why isn't India's G-sec yield risk-free?
India is harder, because the G-sec yield isn't risk-free. Say the 10-year Government of India security yields 6.5% (illustrative, so pull it live from RBI's DBIE tables or CCIL). It's tempting to use it as it is. But Moody's rates the sovereign Baa3 and S&P rates it BBB, after the August 2025 upgrade. Those are the lowest rungs of investment grade. A Baa3 credit's promised yield contains compensation for the possibility it defaults. History has enough sovereign restructurings, even in local currency, that lenders charge for it. Russia in 1998 was one, and Greece in 2012 was one in effect. A "risk-free" rate with default risk inside it fails condition one. Use it anyway and you plant a time bomb, because the ERP build from the risk-and-return work already charges for India risk once, through the CRP. Raw G-sec plus full CRP means sovereign risk gets billed twice. That's about 2 extra points of discount rate, and roughly a fifth of a terminal value vaporized by bookkeeping.
How do you clean the rupee risk-free rate?
Damodaran's clean-INR fix is to subtract the default spread:
Rf(INR, clean) = 10-yr G-sec yield − sovereign default spread (rating- or CDS-based)
= 6.5% − 2.0% = 4.5% (illustrative mid-2026 values)This is the same default spread you used to build the CRP. One number gets used twice, consistently. You subtract it from the government yield to purify the risk-free rate, and you add it back through the equity risk premium, scaled by relative equity volatility and weighted by λ. So sovereign risk is charged once, in the premium, where it belongs. Pull the spread from Damodaran's Country Default Spreads and Risk Premiums page, which is updated each January and revised during the year. You can take either the rating-based lookup or the sovereign CDS quote. Either is fine, as long as you write down which.
There's also a cross-check route, which builds the INR rate from inflation differentials. Fisher gives you an alternative construction: Rf(INR) = (1 + Rf(USD)) × (1 + E[inflation INR]) / (1 + E[inflation USD]) − 1. With 4.2% USD, ~4.5% expected Indian inflation and ~2.3% US inflation, you get 1.042 × 1.045/1.023 − 1 ≈ 6.4%. That disagrees with the subtraction route's 4.5%. Between them, the two methods bracket the truth, and the gap tells you the bond market and the rating agencies are pricing India differently. Or the inflation expectations are off. Damodaran accepts both routes. We standardize on the subtraction method, for continuity with the CRP machinery, and use the differential route as a sanity check. Whichever you pick, write it down. Then apply it to every Indian company you ever value. Consistency beats cleverness.
How does the risk-free rate cap terminal growth?
One consequence reaches forward, so note it while the number is in front of you. This same risk-free rate is the ceiling on terminal growth in every DCF you build off this stack. It becomes a rule in the terminal-value work: perpetual g may not exceed the risk-free rate in your own stack, in the same currency, at the same figure. So the cleaned build here caps terminal g at 4.5%, and a build that keeps the raw 6.5% G-sec caps it at 6.5%. Each is defensible on its own, but you must never cross them. Take the cleaned rate for the denominator and the raw one for the cap, and you've handed growth the sovereign default spread you just refused the discount rate. What about India's nominal GDP growth, 8.9 percent in FY2025-26 and roughly 10 to 11 percent on the long-run average (MoSPI, checked October 2026)? It isn't the cap and never was. That number describes today's economy, and it belongs in the explicit forecast window.
What else follows, and where do you get the data?
Two more consequences deserve a pause.
- A USD valuation of an Indian company is legitimate, and TCS is most naturally forecast in USD-heavy revenue anyway. In that case you use the USD Rf (4.2%) and USD-consistent premiums. Country risk still applies through λ·CRP, though, because switching currency doesn't switch the economy the cash flows are earned in. Currency decides the Rf and the inflation inside your forecasts. Geography, meaning where the cash flows are earned, decides the risk premium. Those are two different dials.
- Never rank rates across currencies. An 11% INR WACC isn't "more demanding" than an 8% USD WACC. Roughly 2 points of the gap is inflation differential, and that same inflation is inflating the INR cash-flow forecasts. Rates and cash flows travel together or not at all.
The live sources, named once, go on your Rf card next to the ERP card from the risk-and-return work. For the US 10-yr, use FRED DGS10. For the 10-yr G-sec, use the RBI Database on the Indian Economy or CCIL (ccilindia.com). Damodaran's ctryprem page gives the default spread. For sovereign ratings, use Moody's/S&P press pages, or that same Damodaran table, which lists them. Date-stamp everything.