Description

The pricing kernel is expected to monotonically decrease under risk aversion, yet empirical studies
consistently document non-monotonic patterns. This phenomenon is called the pricing kernel puz-
zle. (Linn et al., 2018) proposed the Conditional Density Integration (CDI) method to resolve this
puzzle by aligning risk-neutral and physical density information sets. This study critically analyzes
CDI’s performance under structural volatility breaks, for which three volatility jump scenarios are
modeled. The results reveal that while CDI produces monotonic kernels in stable regimes, it fails
under volatility jumps, generating erratic, non-monotonic segments, known under the pricing ker-
nel puzzle. An explanation for this could be the reliance on a global parameter set used for the
estimation, that is based on an averaging process.