How Negative Sovereign Yields Broke the Capital Asset Pricing Model

In September 2011, as the 10-year US Treasury yield slipped below 2.00% for the first time in six decades, New York University finance professor Aswath Damodaran looked at the standard discounted cash flow models used by every major investment bank and realized the math was breaking. The Capital Asset Pricing Model (CAPM) requires a "risk-free rate" to anchor its valuation of riskier assets. But with central banks aggressively buying bonds through quantitative easing, that anchor had been artificially suppressed. Damodaran noted that using a 2.00% base rate would mathematically inflate the value of all equities to absurd levels, while substituting a historical 4.00% average would mean deliberately ignoring the actual cost of capital in the market.[3]

The problem is structural. Every corporate valuation begins with a simple premise: an investor requires a guaranteed return plus a premium for taking on equity risk. For generations, US government debt provided that guaranteed baseline. But as Jason A. Voss pointed out in 2012 for the CFA Institute, the assumption that any sovereign debt is truly devoid of default or purchasing-power risk is a mathematical convenience, not an economic reality. "There is no such thing as a risk-free asset," Voss argued, pointing to the 2011 US credit rating downgrade as proof that even the foundational asset of global finance carries structural risk.[1]

When yields drop below zero—as they did for $18 trillion of global sovereign debt by 2019—the CAPM formula produces nonsensical outputs. The Incwert analysis of negative yield economies demonstrates that if a German bund yields −0.50%, plugging that figure into a standard valuation model implies that investors are paying for the privilege of taking no risk. This artificially compresses the discount rate and sends the theoretical value of long-duration growth stocks toward infinity, divorcing the model from any grounding in corporate fundamentals.[4]

Analysts are forced into a methodological corner. They can use the current manipulated rate, which assumes central bank intervention is permanent, or they can normalize the rate to a historical average, which assumes the current market is wrong. Damodaran's 2011 framework highlighted this exact tension: if an analyst replaces a 2.00% actual yield with a 4.00% normalized yield, they are effectively telling their clients to ignore the actual borrowing costs available in the real economy.[3]

The QuickRead 2026 retrospective on the risk-free rate confirms that this temporary crisis became a permanent methodological fracture. Deconstructing the idea of the risk-free rate reveals that modern finance has been using a proxy that no longer exists. If the US Treasury is subject to inflation risk, political brinkmanship, and central bank manipulation, it is merely the lowest-risk asset, not a risk-free one.[5]

This distinction matters immensely for capital allocation. If a corporate board uses a 1.50% risk-free rate to evaluate a new factory, the project's hurdle rate drops, and marginal investments appear profitable. If they use a 4.00% normalized rate, the same factory is rejected. The disappearance of a true risk-free rate means that capital allocation decisions are now driven as much by an analyst's philosophical view of central banking as by the underlying economics of the business.[3][6]

The strongest counter-argument to abandoning CAPM is that relative valuation models are even worse. While Damodaran explored alternatives to CAPM in April 2011, he conceded that replacing a flawed macroeconomic anchor with peer-based multiples simply imports the market's collective errors into the valuation. If the entire sector is overvalued due to low rates, a relative model will confidently tell you to overpay.[2]

The resolution to this fracture is not a new formula, but a shift in how financial models are interpreted. The risk-free rate can no longer be treated as an objective truth pulled from a Bloomberg terminal. It is a subjective assumption that must be stress-tested. The era of plugging in the 10-year yield and trusting the output is over; the math only works now if the analyst explicitly prices in the risk that the baseline itself is broken.[1][6]