Reverse Engineering in Accounting-Based Equity Valuation Part II: Empirical Evidence on the Errors-in-Expectations Hypothesis

Value investing via Bayesian inference

“…A main takeaway from the studies in the field of value investing is that capturing "value" requires more than one dimension (cheapness and quality), that is, B/P is not a sufficient indicator of value. Despite a large history of research on the relation of B/P and returns however, it remains unclear under which conditions a stock should be seen as cheap: cheapness portfolios are constructed using quantile-based approaches, for example, in Piotroski (2000) or Asness et al (2019), or using absolute boundaries, like in Huefner et al (2021). As a starting point for the model development, we use pricing error research.…”

“…As a result, using book value as a proxy for cheapness is a conservative measure. Huefner et al (2021) extend B/P by incorporating short-term residual income (STE 0 ) into this equation, so that V I,0 equals a RIM with no residual income after the precise forecast horizon, leading to the following slightly upgraded modeling variant of intrinsic value:…”

“…This can lead investors into the so-called value trap, as value stocks may be cheap for a fundamentally justified reason. 4 That is why other contributions refine the B/P strategy by including additional metrics of fundamental performance or different proxies for intrinsic value-like Frankel and Lee (1998), Piotroski (2000), Bird and Gerlach (2003), Mohanram (2005), Piotroski and So (2012), Penman andReggiani (2013, 2018), Lee (2015), , Asness et al (2019), Li and Mohanram (2019), Walkshäusl (2020) and Huefner et al (2021). These studies provide evidence for the predictive ability of fundamental quality for future stock returns on the US market before and after the financial crisis.…”

“…Our analysis primarily anchors on Frankel and Lee (1998), Gao et al (2019), Rueenaufer (2022), and Huefner et al (2021). As alternatives to B/P, Frankel and Lee (1998), Gao et al (2019), and Huefner et al (2021) provide several point-estimate measures for cheapness that form benchmarks for the Bayesian workflow proposed by Rueenaufer (2022). Huefner et al (2021) derive a parsimonious quality index based on valuation theory similar to Piotroski (2000)'s FScore and prove to be empirically valid.…”

“…As alternatives to B/P, Frankel and Lee (1998), Gao et al (2019), and Huefner et al (2021) provide several point-estimate measures for cheapness that form benchmarks for the Bayesian workflow proposed by Rueenaufer (2022). Huefner et al (2021) derive a parsimonious quality index based on valuation theory similar to Piotroski (2000)'s FScore and prove to be empirically valid. This quality index complements the cheapness measures in our analysis.…”