Is the Credit Spread Puzzle a Myth?

Structural models of default can identify asset value dynamics and the location of the default boundary from either (observable) credit spreads or (latent) default probabilities. The latter approach uses historical default rates as proxies, which provide such low statistical power that assumptions regarding asset value dynamics (e.g., geometric Brownian motion) typically go untested. In contrast, calibration via credit spreads offers much more power, and clearly detects jumps in (risk-neutral) asset value dynamics. A large market price of risk on jumps is required to match historical default rates, thus reaffirming a "credit spread puzzle" for investment-grade but not high-yield bonds