The Pricing Kernel Puzzle in Forward Looking Data

Abstract

The pricing kernel puzzle concerns the locally increasing empirical pricing kernel, which is inconsistent with a risk-averse representative investor in a single period, single state variable setting. Some recent papers worry that the puzzle is caused simply by the mismatch of backward looking subjective and forward looking risk-neutral distributions of index returns. By using a novel test and forward looking information only, we generally confirm the existence of a u-shaped pricing kernel puzzle in the S&P 500 options data. The evidence is weaker for tests against an alternative with a risk-neutral investor and for longer horizons.

Improving volatility prediction and option valuation using VIX information: A volatility spillover GARCH model

“…Cuesdeanu and Jackwerth () argue that there exists a nonmonotonic pricing kernel in the S&P 500 option data. Babaoğlu, Christoffersen, Heston, and Jacobs () show that nonmonotonic pricing kernel use has economic benefit and improves option fit by 17% on average.…”

Moment Risk Premiums in Option Markets: On Measurement, Structure, and Investment Implications

“…Upside dispersion swaps for less extreme quantiles either have positive or negligible mean returns and do not contribute to the generally negative overall premiums of dispersion swaps. These results on upside dispersion swaps correspond with the findings by Bakshi et al (2010) for other claims on the upside and the findings by Cuesdeanu and Jackwerth (2018) on a U-shaped pricing kernel. In general, our results are in line with a rationale based on the distinction between good and bad dispersion (Kilic and Shaliastovich 2019).…”

Estimating real‐world probabilities: A forward‐looking behavioral framework

“…">The traditional method to compute the SDF compares a backward‐looking forecast with a forward‐looking RND, questioning whether the pricing kernel puzzle could be caused by misaligned expectations (see Brown & Jackwerth, 2012; Beare, 2011; Beare & Schmidt, 2016; Carr & Wu, 2003; Grith et al, 2017; Yatchew & Härdle, 2006). While several papers address this information mismatch (see Cuesdeanu & Jackwerth, 2018a; Linn et al, 2018; Sala et al, 2016), we are the first to estimate the real‐world pricing kernel, including both risk‐preferences and behavioral effects, in a consistent forward‐looking framework. Previous studies indirectly attribute to sentiment all deviations between a traditional kernel and the empirical SDF.…”